# An improved evaluation of the seismic/geodetic deformation-rate ratio for the Zagros Fold-and-Thrust collisional belt

An improved evaluation of the seismic/geodetic deformation-rate ratio for the Zagros... Summary We present an improved picture of the ongoing crustal deformation field for the Zagros Fold-and-Thrust Belt continental collision zone by using an extensive combination of both novel and published GPS observations. The main results define the significant amount of oblique Arabia–Eurasia convergence currently being absorbed within the Zagros: right-lateral shear along the NW trending Main Recent fault in NW Zagros and accommodated between fold-and-thrust structures and NS right-lateral strike-slip faults on Southern Zagros. In addition, taking into account the 1909–2016 instrumental seismic catalogue, we provide a statistical evaluation of the seismic/geodetic deformation-rate ratio for the area. On Northern Zagros and on the Turkish–Iranian Plateau, a moderate to large fraction (∼49 and >60 per cent, respectively) of the crustal deformation occurs seismically. On the Sanandaj–Sirjan zone, the seismic/geodetic deformation-rate ratio suggests that a small to moderate fraction (<40 per cent) of crustal deformation occurs seismically; locally, the occurrence of large historic earthquakes (M ≥ 6) coupled with the high geodetic deformation, could indicate overdue M ≥ 6 earthquakes. On Southern Zagros, aseismic strain dominates crustal deformation (the ratio ranges in the 15–33 per cent interval). Such aseismic deformation is probably related to the presence of the weak evaporitic Hormuz Formation which allows the occurrence of large aseismic motion on both subhorizontal faults and surfaces of décollement. These results, framed into the seismotectonic framework of the investigated region, confirm that the fold-and-thrust-dominated deformation is driven by buoyancy forces; by contrast, the shear-dominated deformation is primary driven by plate stresses. Creep and deformation, Satellite geodesy, Asia, Earthquake hazards, Seismicity and tectonics 1 INTRODUCTION Geodetic and seismic strain-rate comparison provides significant insights into the seismic hazard of regions subject to relevant tectonic deformation. In particular, when the coseismic displacement is lower than the geodetic displacement across faults, the excess geodetic strain can be released either through large impending earthquakes or in aseismic mode (e.g. Davies et al.1997; Clarke et al.1998). Although this approach has been applied to several regions worldwide, there are several uncertainties regarding the physical significance of the deformation-rates mismatch over varying spatial and temporal scales. Roughly speaking, geodetically observed strains may include both elastic and anelastic components, and in many cases, it is difficult to differentiate the two without a priori knowledge of the rheology of the crust under investigation. Because only the elastic strain is released by earthquakes, the comparison of geodetic and seismic strain-rates may not balance in regions cut by creeping faults or where significant deformation takes place plastically. Thus, verifying actual mismatch between the geodetic and the seismic deformation-rate and estimating its extent require a high level of accuracy and spatiotemporally dense geodetic observations. Apparent imbalance between these proxies of strain-rate may also occur where the seismic catalogue is incomplete, as in the case where the seismic cycle is longer than the duration of the observation period. The Zagros Fold-and-Thrust Belt (ZFTB) continental collision zone is one of the youngest and most seismically active zones on Earth (Fig. 1). By using both instrumental and historical seismicity data coupled with deformation-rates deduced from plate movements, Jackson & McKenzie (1988) made the first quantitative comparison between seismic and deformation-rates of ZFTB. Based on a combined analysis of geodetic and seismic data that had accumulated over the years, Masson et al. (2005) provided a more recent estimation of the contemporaneous seismic versus geodetic deformation-rate ratio. They observed that a low seismic strain-rate, especially in the southern sector of ZFTB, accounted for a small percentage of the observed deformation-rate. This has indicated that the crust is deforming mostly aseismically, or that elastic strain is being accumulated. Figure 1. View largeDownload slide Regional plate tectonic setting of the study region and surrounding areas. Points and hexagons indicate GPS sites used in this work: GPS sites processed in this study are coloured in yellow, while GPS data published by other authors are reported as black symbols (see Table S4, Supporting Information). GPS sites used to estimate the Euler vector for the Arabian plate are reported as hexagons (see the Supporting Information section for details). Plate boundaries are marked as red lines (Bird 2003). Abbreviations are as follows: BS, Black Sea; CS, Caspian Sea; EAF, Eastern Anatolian fault; MS, Mediterranean Sea; SB, Sinai Block; PG, Persian Gulf; GO, Gulf of Oman; AS, Arabian Sea and Me, Mesopotamia. Dashed box shows the study area. The map is plotted in a Mercator projection. Figure 1. View largeDownload slide Regional plate tectonic setting of the study region and surrounding areas. Points and hexagons indicate GPS sites used in this work: GPS sites processed in this study are coloured in yellow, while GPS data published by other authors are reported as black symbols (see Table S4, Supporting Information). GPS sites used to estimate the Euler vector for the Arabian plate are reported as hexagons (see the Supporting Information section for details). Plate boundaries are marked as red lines (Bird 2003). Abbreviations are as follows: BS, Black Sea; CS, Caspian Sea; EAF, Eastern Anatolian fault; MS, Mediterranean Sea; SB, Sinai Block; PG, Persian Gulf; GO, Gulf of Oman; AS, Arabian Sea and Me, Mesopotamia. Dashed box shows the study area. The map is plotted in a Mercator projection. Here, we present an improved picture of the ongoing crustal deformation field of ZFTB, based on an extensive combination of novel observations rigorously integrated with published geodetic velocities. In addition, we compare the GPS (Global Positioning System)-based moment-rates with those derived from earthquake catalogue in order to provide an updated statistical evaluation of the seismic/geodetic deformation-rate ratio for the whole ZFTB. Previous studies, in estimating the seismic/geodetic deformation-rate ratio over the ZFTB, simply used the existing (historical and instrumental) record of seismicity; here we define the long-term moment release rate by adopting a truncated Gutenberg–Richter relation. This method has the advantage of being insensitive to the short duration of the observation period compared with the typical length of the seismic cycle. 2 BACKGROUND SETTING 2.1 Geological and tectonic setting The ZFTB represents the present-day boundary between the Arabian and Eurasian plates (Fig. 1). ZFTB extends for about 2000 km from eastern Turkey where it connects with the Eastern Anatolian Fault (EAF), through SW Iran to the Oman Sea, where it connects to the Makran subduction zone. The belt varies in width (from ∼150 km in the west to ∼300 km in the east) with two main embayments (Kirkuk and Dezful) and two main arcs (Lorestan and Fars; Fig. 2). Figure 2. View largeDownload slide (a) Simplified structural map of the Zagros area derived from Berberian (1995) and Allen et al. (2013), and references therein. A number of faults are inferred from changes of stratigraphic level, and are not directly observed in the field (Berberian 1995). The dashed blue line is a simplified representation of MRF. The 2° × 2° grid (and associated id) used for the geodetic and seismic moment-rate computations is reported as blue lines. Abbreviations are as follows: MRF, Main Recent Fault; MZRF, Main Zagros Reverse Fault; HZF, High Zagros Fault; KhF, Khanaqin Fault, IHF, Izeh–Hendijan Fault; KaF, Kazerun Fault. KaF divides the Zagros into Northern and Southern Zagros. (b) Historical earthquakes (yellow squares) occurring during 800–1908 CE are also reported (Mirzaei et al.1997; Ambraseys & Jackson 1998; see also Table S1 in Supporting Information). The map is plotted in an oblique Mercator projection. Figure 2. View largeDownload slide (a) Simplified structural map of the Zagros area derived from Berberian (1995) and Allen et al. (2013), and references therein. A number of faults are inferred from changes of stratigraphic level, and are not directly observed in the field (Berberian 1995). The dashed blue line is a simplified representation of MRF. The 2° × 2° grid (and associated id) used for the geodetic and seismic moment-rate computations is reported as blue lines. Abbreviations are as follows: MRF, Main Recent Fault; MZRF, Main Zagros Reverse Fault; HZF, High Zagros Fault; KhF, Khanaqin Fault, IHF, Izeh–Hendijan Fault; KaF, Kazerun Fault. KaF divides the Zagros into Northern and Southern Zagros. (b) Historical earthquakes (yellow squares) occurring during 800–1908 CE are also reported (Mirzaei et al.1997; Ambraseys & Jackson 1998; see also Table S1 in Supporting Information). The map is plotted in an oblique Mercator projection. The Main Zagros Reverse Fault (MZRF), juxtaposing the Palaeozoic-Cretaceous stratigraphy against the low-grade metamorphics and Cretaceous cover sequence of the Sanandaj–Sirjan metamorphic belt, is considered as the suture between the Arabian plate and the Iranian Plateau and represents the NE limit of the ZFTB (e.g. Authemayou et al.2006; Paul et al.2006, and references therein). The Main Recent fault (MRF), a young seismically active right-lateral strike-slip fault with a prevailing NW–SE strike, interconnects the MZRF with the EAF termination (Copley & Jackson 2006; Figs 1 and 2). Southwest of the MZRF, the Zagros Mountains are usually divided into the High Zagros (or Imbricate Belt) to the NE and the Simply Folded Zone to the SW. The former, exposing stratigraphic levels in the Mesozoic and Palaeozoic ages, lies between the original suture (i.e. MZRF) and a major thrust, the High Zagros Fault (HZF), which is mapped as running roughly parallel to the suture (Berberian 1995; Bosold et al.2005). The latter, exposing a Mesozoic–Cenozoic (Neotethys) mixed carbonate-clastic succession (Palaeozoic strata are rarely exposed), lies between the HZF, the Persian Gulf and the neighbouring plains of Mesopotamia representing the current flexural foredeep of the ZFTB (e.g. Aqrawi et al.2010, and references therein). In the Fars, these cover sequences are decoupled from the underlying basement along the Lower Cambrian evaporitic Hormuz Formation (1–3 km thick), which has played a significant role in controlling the deformation of the cover (McQuarrie 2004; Molinaro et al.2004; Jahani et al.2009). This formation has emerged as numerous salt diapirs at the surface, and has brought fragments of igneous and metamorphic rocks from the underlying basement (Kent 1979). The evaporitic Hormuz Formation is present across a wide area of the Zagros and Middle East (e.g. Edgell 1991) and contains thick evaporites, mainly halite. However, the lack of diapirs and plugs in the Dezful embayment and Lorestan arc has been attributed to the absence of the Hormuz Formation in these areas (Talbot & Alavi 1996). Moreover, the presence of a number of post-Triassic evaporitic or shale layers has been recognized in different regions of the belt and has been interpreted as possible intermediate décollement levels within the sedimentary cover (Sherkati et al.2006). The ZFTB involves parallel folds that have formed by buckling above the evaporitic Hormuz Formation (Colman-Sadd 1978; McQuarrie 2004; Mouthereau et al.2007), thrust-cored anticlines (Sattarzadeh et al.1999; Molinaro et al.2004), and thrusts verging both to the SW and to the NE (the last category responsible for some recent large earthquakes; e.g. Nissen et al.2014; Elliott et al.2015) in the sedimentary cover and its underlying basement, implying both thin- and thick-skinned shortening across the belt (Talbot & Alavi 1996; Mouthereau et al.2007). The N-S Arabia–Eurasia convergence process seems transformed by some roughly N-S-trending strike-slip faults (e.g. the Izeh−Hendijan, Kazerun and Khanaqin right-lateral faults; Fig. 2a). Among these faults, the Kazerun fault constitutes the western limit of the evaporitic Hormuz basin (Edgell 1991). The timing of initial Arabia–Eurasia collision is still debated, with recently published estimates spanning from the Late Cretaceous/Palaeocene (Ghasemi & Talbot 2006; Mazhari et al.2009), the Early Miocene (Okay et al.2010) to the Middle/Late Miocene (Guest et al.2006). Moreover, opinions are now converging to a Latest Eocene (∼35 Ma) age for initial collision (when the distal Arabian continental margin, driven by its negative buoyancy, was underthrusted beneath the upper Iranian block) and to a Late Oligocene (∼25 Ma) age for the onset of the continental crustal thickening (see Mouthereau et al.2012, and references therein for a review). Palaeozoological evidence suggests that the Arabian plate land bridge for large mammal migration between India and Africa initiated only as late as 17 Ma (e.g. Tchernov et al.1987; Goldsmith et al.1994). 2.2 Historical seismicity Available historical seismic catalogues for the Zagros area document the occurrence of large (M > 6) earthquakes since AD 200. However, the accuracy of these catalogues is non-uniform and questionable mainly due to the sparsely populated area (e.g. Mirzaei et al.1997; Ambraseys & Jackson 1998, and references therein). Studies focusing on the completeness of available data in the study area and surrounding regions (Ambraseys & Melville 1982; Melville 1984; Ambraseys 1989) have pointed out that many small and moderate earthquakes have most likely been overlooked and that available catalogues can be reasonably considered complete for earthquakes with M ≥ 6 since 1860 (Mirzaei et al.1997). Historic earthquakes striking the area during the last millennium are largely concentrated on Southern Zagros and along the Turkish–Iranian Plateau (Fig. 2b; Mirzaei et al.1997; Ambraseys & Jackson 1998), where the MRF connects with the EAF termination across an 80-km-wider shear zone. Such a shear zone is made by numerous subparallel NW–SE striking faults characterized by right-lateral strike-slip kinematics (e.g. Copley & Jackson 2006). In detail, the stronger historic earthquakes (M ≥ 7.0; AD 1008, AD 1042, AD 1696, AD 1721, AD 1780, AD 1840, AD 1903; see also Table S1, Supporting Information section for details) are mainly concentrated on the Turkish–Iranian Plateau, while the Southern Zagros have experienced only a single large earthquake (estimated magnitude M = 7.1) in AD 1440. 3 GEODETIC AND INSTRUMENTAL SEISMIC DATA 3.1 GPS data Here, we analysed an extensive GPS data set covering ∼15 yr of observations, from 1999.00 up to 2014.00, and covering the Arabian plate and surrounding areas. GPS data come from public online archives, which include CDDIS (https://cddis.nasa.gov), EUREF (www.epncb.oma.be), SOPAC (http://sopac.ucsd.edu/), UNAVCO (www.unavco.org) and NGS (http://geodesy.noaa.gov). Raw GPS data were processed by using the GAMIT/GLOBK software (Herring et al.2010; www-gpsg.mit.edu), by adopting the strategy described in Palano (2015). By using the GLORG module of GLOBK, all the GAMIT solutions and their full covariance matrices were combined to estimate a consistent set of positions and velocities in the ITRF2008 reference frame. Our analysed network does not cover entirely the Zagros area. Nonetheless, since our solution shares some stations with the ones processed by Hessami et al. (2006), Reilinger et al. (2006), Walpersdorf et al. (2006, 2014), Masson et al. (2007), Tavakoli et al. (2008), Peyret et al. (2009), Djamour et al. (2011), Mousavi et al. (2013), Palano et al. (2013) and Zarifi et al. (2013), we rigorously integrated data sets by applying a Helmert transformation of the different estimated velocity fields, in order to generate a combined velocity field aligned to a unified frame such as the ITRF2008 one. To adequately show the crustal deformation pattern over the area investigated, we rotate the unified ITRF2008 GPS velocity solution into an Arabian fixed reference frame (Fig. 3a; see Tables S2–S4, Supporting Information section for additional details). Figure 3. View largeDownload slide (a) GPS velocities and 95 per cent confidence ellipses in a fixed Arabian plate (see Table S2 and S3, Supporting Information section for details). Abbreviations are as Fig. 2. The dashed blue line is a simplified representation of MRF. The dashed dark-blue lines represent the expected Arabia-Eurasia relative motion direction. (b) Geodetic strain-rate field and associate uncertainties: arrows represent the greatest extensional (red) and contractional (blue) horizontal strain-rates. Strain-rates derived from the velocity interpolation only (the grid cell contain zero or just one GPS site) are reported as white colour. The blue lines defined the 2° × 2° grid used for the geodetic and seismic moment-rates computation. Yellow dots stand for GPS sites. The map is plotted in an oblique Mercator projection. Figure 3. View largeDownload slide (a) GPS velocities and 95 per cent confidence ellipses in a fixed Arabian plate (see Table S2 and S3, Supporting Information section for details). Abbreviations are as Fig. 2. The dashed blue line is a simplified representation of MRF. The dashed dark-blue lines represent the expected Arabia-Eurasia relative motion direction. (b) Geodetic strain-rate field and associate uncertainties: arrows represent the greatest extensional (red) and contractional (blue) horizontal strain-rates. Strain-rates derived from the velocity interpolation only (the grid cell contain zero or just one GPS site) are reported as white colour. The blue lines defined the 2° × 2° grid used for the geodetic and seismic moment-rates computation. Yellow dots stand for GPS sites. The map is plotted in an oblique Mercator projection. In addition, by taking into account the observed horizontal velocity field and associated uncertainties, we derived a continuous velocity gradient over the study area on a regular 1° × 1° grid (whose nodes do not coincide with any of the GPS stations) using a ‘spline in tension’ function (Wessel & Bercovici 1998). The 1° value corresponds to the average distance between GPS stations (in degrees). Sites biased by large velocity uncertainties and/or showing suspicious movements with respect to nearby sites (∼1 per cent of the data set here analysed) were not used for the velocity gradient computation. The tension is controlled by a factor T, where T = 0 leads to a minimum curvature (natural bicubic spline), while T = 1 leads to a maximum curvature, allowing for maxima and minima only at observation points (Smith & Wessel 1990). We set T to the value of 0.5, because it represents the optimal value to minimize short wavelength noise (see Palano 2015, and references therein). Lastly, we computed the average 2-D strain-rate tensor (and its standard error) as derivative of the velocities at the centre of each cell. The estimated strain-rates are shown in Fig. 3(b) as principal extensional ($${\dot{\varepsilon }_{H\max }}$$) and shortening ($${\dot{\varepsilon }_{h\min }}$$) horizontal strain-rates. Various methods can be used to derive GPS strain-rates, ranging from simple Delaunay triangulations to more complex parametric inversions (e.g. Haines & Holt 1993), with significant variations in their results. Therefore, since the spatial distribution of our velocity field data is heterogeneous, in order to assess the first-order variability in our strain-rate analysis and the derived moment-rates, we undertook some additional computations (see the Supporting Information section) by adopting the method described in Shen et al. (2015). This method, in interpolating the displacement field, introduces the spatial weighting function of data in various forms (e.g. uniform Gaussian or quadratic spatial weighting function), allowing to obtain finer resolution especially for regions in which data are sparsely distributed. It must be noted that some of the cells contain nil or just a single GPS site and the strain-rates information derived from the velocity interpolation only; because these estimates cannot be considered accurate for the area due to the lack of spatial constraints from data, in the following, information coming from these cells have been omitted. Overall, the GPS-based velocity field (referred to an Arabian fixed reference frame) clearly depicts a clockwise rotation, passing from an SE-ward motion along the Turkish–Iranian Plateau to an SW-directed motion on the Fars Arc. This rotation implies that in the former, the velocity field shows an oblique relationship (more than 25°) in comparison with the predicted convergent motion of Arabia with respect to Eurasia plate, while in the latter the velocity field is near parallel to such a predicted motion (Fig. 3a). Because the trend of the Zagros does not change along strike, both the above-mentioned features imply that shortening across the northwestern sector contributes less to that overall Eurasia–Arabia convergence than does convergence across the southeastern sector, as evidenced by previous studies (e.g. Jackson 1992; Copley & Jackson 2006; Reilinger et al.2006). Descriptions refining these main features are detailed below. Stations located along the Turkish–Iranian Plateau and along the NE border of ZFTB show a small clockwise rotation passing from an SE-directed motion near the Urmia Lake region (Fig. 3a rates of ∼11–12 mm yr−1) to an SSE-directed motion (Fig. 3a rates of ∼9–11 mm yr−1) close to the Qom region. Conversely, stations located in the region spanning the Kirkuk embayment, the Lorestan arc, the MRF (Fig. 3a) and westward the ZFTB frontal area, in the stable Arabian plate (e.g. ISBA, ISSD and ISNA stations), show negligible deformation. Considering the different motion occurring across MRF, a general right-lateral shear of this area, can be deduced. The lack of GPS measurements close to the MRF does not allow direct measurement of the geodetic slip-rate on this fault system; however, considering the pattern previously described we suggest that it accounts for ∼8–11 mm yr−1 of right-lateral slip. The regular stations distribution across the southern sector of ZFTB allows better detecting the crustal deformation field. Westward of MZRF, the ground deformation field is characterized by a prevailing S-directed motion, passing to an SW-ward motion approaching the frontal trust belt. This rotation is coupled by a progressive reduction of geodetic rates passing from values of ∼10–12 to ∼1–3 mm yr−1, respectively, from the inner side to the external side of ZFTB. By using a simple vectorial decomposition (by grouping GPS velocities on each side of a selected fault), we estimate ∼2.8 and ∼3.4 mm yr−1 of right-lateral motion for the Izeh−Hendijan and Kazerun strike-slip faults, respectively. The prevailing contractional nature of ZFTB is evident on the 2-D geodetic strain-rate map (Fig. 3b). In particular, the maximum contractional horizontal strain-rate shows a fan-shaped feature across the southern sector of ZFTB, maintaining always an orthogonal orientation with respect to the curvature of the collisional mountain belt (‘orocline’); across this area, a shortening up to ∼50 nanostrain per year can be recognized. Moving toward the Turkish–Iranian Plateau, such a prevailing contractional pattern progressively turn into a pattern characterized by an NS shortening up to ∼40 nanostrain per year coupled with an EW extension up to ∼45 nanostrainper year, suggesting a prevailing strike-slip deformation of the Plateau. 3.2 Seismic data Instrumental seismic monitoring of the study area started in 1957 when the Iranian Seismological Centre (http://irsc.ut.ac.ir) deployed a seismograph network. Since then, other regional (e.g. IIEES, http://www.iiees.ac.ir and BHRC, http://www.bhrc.ac.ir) and local seismic networks have been established, thereby increasing the number of seismic stations and improving the quality of earthquake observables (i.e. hypocentral locations and source mechanism determinations). From the International Seismological Centre (ISC, www.isc.ac.uk) online catalogue, we compiled a database of more than 66 000 seismic events occurring since 1909 to date and having focal depth ≤ 50 km and M ≥ 1.0 (Fig. 4a; see also Table S5, Supporting Information section). The ISC routinely produce catalogues of earthquake hypocentre locations relying on data contributed by seismological agencies from around the world. The location provided by the ISC is based on P-wave traveltime tables derived from global 1-D earth velocity model (e.g. the ak135 model; Kennett et al. 1995). The resulting hypocentral parameters are based entirely on reported first-arriving P-wave times, which for most events do not include P-wave arrivals corresponding to upgoing ray paths. Therefore, many ISC hypocentres are poorly constrained in focal depth and must be interpreted with caution. Figure 4. View largeDownload slide (a) Instrumental crustal seismicity (M ≥ 2.5) occurring in the investigated area since 1900 (International Seismological Centre, www.isc.ac.uk). (b) Lower hemisphere, equal area projection for FPSs (with M ≥ 5) compiled from the GCMT catalogue (http://www.globalcmt.org; Dziewonski et al.1981; Ekström et al.2012); FPSs are coloured according to rake: red indicates pure thrust faulting, blue is pure normal faulting and yellow is strike-slip faulting. The inset shows a ternary plot of FPSs: each point is plotted based on the plunge of the P, T and B axes of FPS (Frohlich 1992). The dashed lines divide the diagram into faulting styles based on definitions by Zoback (1992): NF is normal faulting, NS is normal and strike-slip faulting, SS is strike-slip faulting, TS is thrust and strike-slip faulting, TF is thrust faulting and U is undefined faulting. The map is plotted in an oblique Mercator projection. Figure 4. View largeDownload slide (a) Instrumental crustal seismicity (M ≥ 2.5) occurring in the investigated area since 1900 (International Seismological Centre, www.isc.ac.uk). (b) Lower hemisphere, equal area projection for FPSs (with M ≥ 5) compiled from the GCMT catalogue (http://www.globalcmt.org; Dziewonski et al.1981; Ekström et al.2012); FPSs are coloured according to rake: red indicates pure thrust faulting, blue is pure normal faulting and yellow is strike-slip faulting. The inset shows a ternary plot of FPSs: each point is plotted based on the plunge of the P, T and B axes of FPS (Frohlich 1992). The dashed lines divide the diagram into faulting styles based on definitions by Zoback (1992): NF is normal faulting, NS is normal and strike-slip faulting, SS is strike-slip faulting, TS is thrust and strike-slip faulting, TF is thrust faulting and U is undefined faulting. The map is plotted in an oblique Mercator projection. Keeping in mind this main limitation, the instrumental seismicity is mainly distributed on the Turkish–Iranian Plateau and along the Zagros collisional belt (Fig. 4a). In the former, seismicity appears rather spread out, while along the latter, seismicity is confined between the Persian Gulf and the MZRF. The area separating these two regions shows a rather low level of seismicity with sporadic occurrence of large and destructive earthquakes (e.g. the M = 7.2 1962 Buyin Zara earthquake; Fig. 4a), while along the Zagros collisional belt the number of moderate size earthquakes (4.5 ≤ M ≤ 6) is large and no earthquakes with magnitude larger than 7 have been recorded. In order to depict the main seismotectonic features of the investigated area, we compiled a database of focal plane solutions (FPSs) from the Global Centroid Moment Tensor online catalogue (Dziewonski et al.1981; Ekström et al.2012; www.globalcmt.org). FPSs with reverse faulting predominate in Southern Zagros and are characterized by strikes subparallel to the local trend of the topography and fold axes (Fig. 4b). In this area, a number of strike-slip solutions are also evident, mostly associated with the N-S-trending strike-slip right-lateral Kazerun fault and along secondary structures within the Simply Folded Zone. Along Northern Zagros, FPSs are mainly distributed along the external front of the collisional belt and are associated with a few strike-slip solutions occurring along MRF (Fig. 4b). The Turkish–Iranian Plateau is characterized by an equal mixture of reverse and strike-slip solutions. Overall, the FPSs patterns clearly depict the prevailing contractional nature of Southern Zagros (locally segmented along N-S-trending strike-slip faults), which passes to a transpressional faulting regime toward the NW. 4 MOMENT-RATES COMPUTATION AND COMPARISONS In the following, we estimated the scalar geodetic and seismic moment-rates in order to provide an improved evaluation of the seismic/geodetic deformation-rate comparison for the ZFTB. To this aim, we divided the investigated regions into regular 2° × 2° grid (Fig. 3b) in order to have a consistent seismic data set (e.g. catalogue duration and range of magnitude) for each cell. Therefore, in the following we refer to this 2° × 2° grid. The geodetic moment-rate (and its standard error) was estimated for cells of surface area A by adopting the Savage & Simpson (1997) formulation:   $${\dot{M}_{{\rm{geod}}}} = 2\mu {H_s}A[{\rm{Max}}(|{\varepsilon _{H\max }}|,|{\varepsilon _{h\min }}|,|{\varepsilon _{H\max }} + {\varepsilon _{h\min }}|)]$$ (1) where μ is the shear modulus of the rocks (taken here as 3.3·× 1010 N m−2), Hs is the seismogenic thickness over which strains accumulate and its elastic part release in earthquakes, εHmax and εhmin are the principal horizontal strain-rates previously described and Max is a function returning the largest of the arguments. The moment-rate estimate from geodetic strain-rates is proportional to the chosen seismogenic thickness Hs. As mentioned earlier, focal depths for many events in the ISC catalogue cannot be determined with sufficient accuracy. Concerning the ZFTB, uncertainties associated to focal depths are on the order of 10 km, even for the ‘best case’ examples where depth phases (pP, sP and sS) are picked (see Engdahl et al.2006 for additional details). Moreover, a number of recent studies, based on local recordings and teleseismic body-waveform modeling have demonstrated that most of the well-located earthquakes occurs in the 10–14 km depth interval and that most of the events deeper than ∼18 km occur in the south-eastern sector of ZFTB (e.g. Talebian & Jackson 2004; Engdahl et al.2006; Nissen et al.2011; Ansari & Zamari 2014). Hence, based on these considerations, we set Hs= 15 km. For each cell, the seismic moment-rate was calculated according to the Hyndman & Weichert (1983) formulation:   $${\dot{M}_{{\rm{seis}}}} = \phi \frac{b}{{\left( {c - b} \right)}}{10^{\left[ {\left( {c - b} \right){M_x} + a + d} \right]}}$$ (2)which is obtained by integrating the cumulative truncated Gutenberg–Richter distribution up to a maximum magnitude Mx, that is, the magnitude of the largest earthquake that could occur within a specified region. φ is a correction for the stochastic magnitude–moment relation; according to Hyndman & Weichert (1983), we assumed φ = 1.27, reflecting a standard error of 0.2 on magnitudes. c and d are the coefficients of the magnitude (M)–scalar moment (Mseis) relation:   $$\log {M_{{\rm{seis}}}} = cM + d$$ (3) According to Hanks & Kanamori (1979), we set c = 1.5 and d = 9.05. We are aware that the earthquake magnitudes in the ISC catalogue refer to different scales (e.g. mb, body wave magnitude; Ms, surface wave magnitude and MD duration magnitude) which ideally should be converted into moment magnitude (Mw) and use that as a standard, given the limitations of the other magnitude scales. Although Karimiparidari et al. (2013) provide some relationships between Mw and the other magnitude types, we prefer to convert all earthquake magnitudes directly into scalar moments by using the above generalized relation, because in any case, both estimations will always suffer from substantial uncertainties. The coefficients a and b are the seismicity level and the slope of the Gutenberg–Richter recurrence relation, respectively:   $$\log {N_M} = a - bM$$ (4) where, for each cell, NM is the cumulative number of earthquakes of magnitude M and larger. In order to estimate a and b, we adopted the maximum likelihood method (Weichert 1980). Both coefficients have been primarily constrained by the small- to mid-size earthquakes (M < 5), using as input the instrumental catalogue. In some few cells (Z008 and Z017), externally located with respect to ZFTB, the number of earthquakes is relatively small (<50) and can be regarded as a poor constraint, especially on the b value estimation. Hence, results obtained for these cells have been omitted. Results achieved for cells containing lesser than 200 earthquakes (Z006, Z016, Z021, Z022, Z024 and Z030) can be viewed as adequate, while those achieved for cells containing more than 200 earthquakes can be considered from good to highly reliable. Estimated parameters are reported in Table S6 of the Supporting Information section. According to eq. (2), the seismic moment-rate estimation for each cell depends by the Mx value. A simple method of calculating Mx is to use the largest earthquake in the historical catalogue and add 0.5 (e.g. Kijko & Graham 1998), however such a method is very limited where there is no significant historical record. Another method is to use scaling relations between the length of the fault and the maximum earthquake (e.g. Wells & Coppersmith 1994). This method can be applied where there are no historical data, but a number of issues come with deciding on whether, and how, to divide the fault up into segments. As alternative, Mx can be estimated by using statistical approaches (see also the Supporting Information section). In this study, we estimated the Mx value by using the MMAX toolbox developed by Kijko & Singh (2011). Such a toolbox, by adopting a wide spectrum of statistical procedures, allows to estimate the Mx values for a given area, in different circumstances (completeness and temporal length of the catalogue, magnitude distribution and uncertainties, number of earthquakes, etc; see Kijko & Singh 2011 for details). For each cell, we estimated the Mx values according with the 12 procedures implemented in the toolbox and using as input two different data sets. In a former step, we used as input the instrumental catalogue which spans the 1909–2016 time interval; the main results are reported in Table S7 in the Supporting Information. In this step, a standard error of 0.25 in magnitude determination of all events have been considered. In a latter step, we used as input a catalogue including both the instrumental catalogue used in the former computation and the historical seismic records (Mirzaei et al.1997; Ambraseys & Jackson 1998); the main results are reported in Table S8 in the Supporting Information. Since magnitude uncertainties in the historical catalogue are hard to measure and quantify (and usually increase further back in time) in this step, a standard error of 0.35 in magnitude determination of all events have been considered. Considering all the results coming from the two steps, the estimations based only on the instrumental data usually underestimate the Mx values, since the ‘seismic cycle’ over the investigated area is longer than the duration of the instrumental record time span. On the other hand, the estimations based on the mixed catalogue (covering the last 1200 yr) agree with the highest observed magnitudes, reflecting however, a strictly dependence of Mx estimations to the historical records (e.g. the number of historic events over higher magnitude ranges in each cell, event magnitude uncertainties). Both these features can be clearly observed from the cumulative frequency–magnitude plots reported in Fig. S6 in the Supporting Information: for instance, the estimated 7.84 Mx value (Tables S8, Supporting Information; the estimation is based on two M = 7.7 earthquakes, whose estimated magnitude are affected by large uncertainties) for cell Z006, is hardly derivable from the plot, while a value of 5.8 (Table S7, Supporting Information; no earthquakes with M > 6 occurred in the cell during the instrumental period) should be more appropriate. Furthermore, it must be also noted that, the Mx value has a significant effect on the seismic moment-rates estimation (2); an increase of 1 mag unit leads to an increase of the moment-rate by a factor of ∼5.4 (e.g. Mazzotti et al.2011). Keeping in mind all these aspects, in the following, we refer to the estimations based on the mixed catalogue. The toolbox does not indicate which of the resulting estimations is the more appropriate; hence, for each cell we selected the one having the smallest uncertainties (see the Supporting Information section and Fig. S7 for details). Finally, for each cell, the seismic moment-rates (and associated uncertainties) have been estimated according to the relationship (2), by taking into account all the parameters as previously defined/computed. Uncertainties in our estimates have been provided as a median value along with an associated 67 per cent confidence interval, by adopting a logic tree approach, similar to the one described in Mazzotti & Adams (2005). The seismic moment-rate can be estimated also by adopting the Kostrov (1974) formulation:   $${\dot{M}_{{\rm{seis}}}} = \frac{1}{{\Delta T}}\sum\limits_{n = 1}^N {M_{{\rm{seis}}}^{(n)}}$$ (5) where N is the number of events occurring in the volume A · Hs (A and Hs represent the surface area and the seismogenic thickness previously defined, respectively), Mseis is the seismic moment of the nth earthquake from the N total earthquakes and ΔT represents the time interval considered. We mainly focused on the results obtained from the truncated Gutenberg–Richter distribution, however, in order to test sensitivity, we additional computed the Kostrov summation (see the Supporting Information section for details). The estimated geodetic and seismic moment-rates are both reported in Fig. 5 and Table 1. Regarding the geodetic moment-rates, we performed an additional computation by taking into account the strain-rate field estimated with the method described in Shen et al. (2015) and reported in Fig. S1a in the Supporting Information. Both geodetic moment-rate estimates (reported in Fig. S8, Supporting Information), with the exception of very few cells (e.g. Z002, Z015, Z020 and Z029), well overlaps within their associated uncertainties. This fact clearly highlights as the crustal pattern of deformation over the investigated area is adequately sampled by both methods, although they rely on different approaches in velocity gradient computation. The following considerations and computations are primarily based on the results coming from the ‘spline in tension’ method, with references to the alternative results where they are significant. Figure 5. View largeDownload slide Estimated moment-rates from (a) geodetic data, (b) instrumental seismicity, by considering the truncated Gutenberg–Richter distribution and (c) instrumental seismicity, by considering the cumulative Kostrov summation method (see Table 1 for details). The maps are plotted in an oblique Mercator projection. Figure 5. View largeDownload slide Estimated moment-rates from (a) geodetic data, (b) instrumental seismicity, by considering the truncated Gutenberg–Richter distribution and (c) instrumental seismicity, by considering the cumulative Kostrov summation method (see Table 1 for details). The maps are plotted in an oblique Mercator projection. Table 1. Moment-rates (N·m yr−1) estimated from geodetic data (GM1) and instrumental seismicity estimated by integrating the cumulative truncated Gutenberg–Richter distribution up to a maximum magnitude Mx (SM–GR) and by adopting the cumulative Kostrov summation method (SM–K). Cell ID  Long.  Lat.  GM1  SM–GR  SM–K  SCC–GR  SCC–K        (1017 N·m yr−1)  (1017 N·m yr−1)  (1017 N·m yr−1)  (per cent)  (per cent)  Z001  43.000  39.000  16.2 ± 1.2  $$13.1_{ - 4.1}^{ + 7.1}$$  3.1  $$80.7_{ - 20.7}^{ + 35.1}$$  18.9  Z002  45.000  39.000  17.5 ± 1.4  $$5.3_{ - 2.4}^{ + 6.6}$$  9.0  $$30.1_{ - 11.1}^{ + 30.0}$$  51.6  Z005  45.000  37.000  12.1 ± 2.2  $$2.7_{ - 1.2}^{ + 3.4}$$  0.1  $$22.2_{ - 8.8}^{ + 22.3}$$  0.8  Z006  47.000  37.000  7.4 ± 0.6  $$6.6_{ - 2.6}^{ + 5.7}$$  0.01  $$88.6_{ - 28.8}^{ + 62.3}$$  0.6  Z007  49.000  37.000  6.3 ± 2.1  $$8.7_{ - 4.3}^{ + 9.9}$$  1.1  $$138.3_{ - 66.1}^{ + 132.0}$$  16.8  Z011  49.000  35.000  7.2 ± 1.7  $$6.6_{ - 2.8}^{ + 8.2}$$  12.9  $$91.2_{ - 39.3}^{ + 106.1}$$  179.4  Z012  51.000  35.000  8.64 ± 2.1  $$2.4_{ - 1.3}^{ + 4.5}$$  0.1  $$27.4_{ - 12.7}^{ + 42.2}$$  1.6  Z013  45.000  33.000  5.0 ± 1.6  $$1.6_{ - 0.7}^{ + 1.8}$$  0.1  $$31.4_{ - 13.8}^{ + 30.1}$$  2.1  Z014  47.000  33.000  7.8 ± 3.1  $$6.3_{ - 2.5}^{ + 5.4}$$  1.0  $$81.2_{ - 35.9}^{ + 60.9}$$  12.2  Z015  49.000  33.000  12.1 ± 2.0  $$2.5_{ - 1.0}^{ + 2.2}$$  0.8  $$20.5_{ - 7.2}^{ + 15.0}$$  6.8  Z016  51.000  33.000  6.9 ± 2.2  $$0.7_{ - 0.4}^{ + 2.2}$$  0.02  $$9.6_{ - 5.2}^{ + 26.0}$$  0.3  Z019  49.000  31.000  9.4 ± 2.6  $$5.5_{ - 2.1}^{ + 4.5}$$  0.2  $$58.2_{ - 22.0}^{ + 39.9}$$  2.6  Z020  51.000  31.000  12.5 ± 2.1  $$4.5_{ - 2.3}^{ + 7.6}$$  1.0  $$35.9_{ - 15.2}^{ + 48.7}$$  7.6  Z021  53.000  31.000  4.8 ± 2.6  $$1.8_{ - 0.9}^{ + 3.4}$$  0.1  $$37.8_{ - 23.1}^{ + 66.0}$$  1.1  Z022  55.000  31.000  3.4 ± 1.9  $$0.1_{ - 0.0}^{ + 0.1}$$  0.3  $$1.7_{ - 1.0}^{ + 3.2}$$  8.4  Z023  57.000  31.000  6.9 ± 1.6  $$3.2_{ - 1.2}^{ + 2.5}$$  0.7  $$46.9_{ - 16.8}^{ + 30.5}$$  10.8  Z025  51.000  29.000  13.7 ± 3.0  $$2.9_{ - 1.0}^{ + 1.9}$$  2.2  $$21.4_{ - 6.9}^{ + 11.8}$$  15.7  Z026  53.000  29.000  17.1 ± 1.3  $$5.7_{ - 3.3}^{ + 16.4}$$  1.1  $$33.3_{ - 15.4}^{ + 76.3}$$  6.1  Z027  55.000  29.000  12.9 ± 2.0  $$0.9_{ - 0.4}^{ + 1.4}$$  0.7  $$7.1_{ - 2.9}^{ + 8.6}$$  5.5  Z028  57.000  29.000  13.8 ± 2.1  $$4.6_{2.1}^{ + 6.0}$$  2.6  $$33.5_{ - 13.0}^{ + 35.1}$$  18.5  Z029  59.000  29.000  6.2 ± 0.6  $$1.9_{ - 1.1}^{ + 5.1}$$  0.7  $$31.3_{ - 14.1}^{ + 65.6}$$  11.4  Z031  53.000  27.000  15.4 ± 2.7  $$2.3_{ - 1.0}^{ + 2.3}$$  0.6  $$15.1_{ - 5.3}^{ + 11.9}$$  3.7  Z032  55.000  27.000  16.9 ± 0.7  $$4.7_{ - 2.0}^{ + 5.3}$$  2.4  $$27.5_{ - 9.6}^{ + 25.1}$$  14.5  Z033  57.000  27.000  21.6 ± 3.4  $$4.4_{ - 1.6}^{ + 3.1}$$  1.7  $$20.5_{ - 6.3}^{ + 11.9}$$  7.8  Cell ID  Long.  Lat.  GM1  SM–GR  SM–K  SCC–GR  SCC–K        (1017 N·m yr−1)  (1017 N·m yr−1)  (1017 N·m yr−1)  (per cent)  (per cent)  Z001  43.000  39.000  16.2 ± 1.2  $$13.1_{ - 4.1}^{ + 7.1}$$  3.1  $$80.7_{ - 20.7}^{ + 35.1}$$  18.9  Z002  45.000  39.000  17.5 ± 1.4  $$5.3_{ - 2.4}^{ + 6.6}$$  9.0  $$30.1_{ - 11.1}^{ + 30.0}$$  51.6  Z005  45.000  37.000  12.1 ± 2.2  $$2.7_{ - 1.2}^{ + 3.4}$$  0.1  $$22.2_{ - 8.8}^{ + 22.3}$$  0.8  Z006  47.000  37.000  7.4 ± 0.6  $$6.6_{ - 2.6}^{ + 5.7}$$  0.01  $$88.6_{ - 28.8}^{ + 62.3}$$  0.6  Z007  49.000  37.000  6.3 ± 2.1  $$8.7_{ - 4.3}^{ + 9.9}$$  1.1  $$138.3_{ - 66.1}^{ + 132.0}$$  16.8  Z011  49.000  35.000  7.2 ± 1.7  $$6.6_{ - 2.8}^{ + 8.2}$$  12.9  $$91.2_{ - 39.3}^{ + 106.1}$$  179.4  Z012  51.000  35.000  8.64 ± 2.1  $$2.4_{ - 1.3}^{ + 4.5}$$  0.1  $$27.4_{ - 12.7}^{ + 42.2}$$  1.6  Z013  45.000  33.000  5.0 ± 1.6  $$1.6_{ - 0.7}^{ + 1.8}$$  0.1  $$31.4_{ - 13.8}^{ + 30.1}$$  2.1  Z014  47.000  33.000  7.8 ± 3.1  $$6.3_{ - 2.5}^{ + 5.4}$$  1.0  $$81.2_{ - 35.9}^{ + 60.9}$$  12.2  Z015  49.000  33.000  12.1 ± 2.0  $$2.5_{ - 1.0}^{ + 2.2}$$  0.8  $$20.5_{ - 7.2}^{ + 15.0}$$  6.8  Z016  51.000  33.000  6.9 ± 2.2  $$0.7_{ - 0.4}^{ + 2.2}$$  0.02  $$9.6_{ - 5.2}^{ + 26.0}$$  0.3  Z019  49.000  31.000  9.4 ± 2.6  $$5.5_{ - 2.1}^{ + 4.5}$$  0.2  $$58.2_{ - 22.0}^{ + 39.9}$$  2.6  Z020  51.000  31.000  12.5 ± 2.1  $$4.5_{ - 2.3}^{ + 7.6}$$  1.0  $$35.9_{ - 15.2}^{ + 48.7}$$  7.6  Z021  53.000  31.000  4.8 ± 2.6  $$1.8_{ - 0.9}^{ + 3.4}$$  0.1  $$37.8_{ - 23.1}^{ + 66.0}$$  1.1  Z022  55.000  31.000  3.4 ± 1.9  $$0.1_{ - 0.0}^{ + 0.1}$$  0.3  $$1.7_{ - 1.0}^{ + 3.2}$$  8.4  Z023  57.000  31.000  6.9 ± 1.6  $$3.2_{ - 1.2}^{ + 2.5}$$  0.7  $$46.9_{ - 16.8}^{ + 30.5}$$  10.8  Z025  51.000  29.000  13.7 ± 3.0  $$2.9_{ - 1.0}^{ + 1.9}$$  2.2  $$21.4_{ - 6.9}^{ + 11.8}$$  15.7  Z026  53.000  29.000  17.1 ± 1.3  $$5.7_{ - 3.3}^{ + 16.4}$$  1.1  $$33.3_{ - 15.4}^{ + 76.3}$$  6.1  Z027  55.000  29.000  12.9 ± 2.0  $$0.9_{ - 0.4}^{ + 1.4}$$  0.7  $$7.1_{ - 2.9}^{ + 8.6}$$  5.5  Z028  57.000  29.000  13.8 ± 2.1  $$4.6_{2.1}^{ + 6.0}$$  2.6  $$33.5_{ - 13.0}^{ + 35.1}$$  18.5  Z029  59.000  29.000  6.2 ± 0.6  $$1.9_{ - 1.1}^{ + 5.1}$$  0.7  $$31.3_{ - 14.1}^{ + 65.6}$$  11.4  Z031  53.000  27.000  15.4 ± 2.7  $$2.3_{ - 1.0}^{ + 2.3}$$  0.6  $$15.1_{ - 5.3}^{ + 11.9}$$  3.7  Z032  55.000  27.000  16.9 ± 0.7  $$4.7_{ - 2.0}^{ + 5.3}$$  2.4  $$27.5_{ - 9.6}^{ + 25.1}$$  14.5  Z033  57.000  27.000  21.6 ± 3.4  $$4.4_{ - 1.6}^{ + 3.1}$$  1.7  $$20.5_{ - 6.3}^{ + 11.9}$$  7.8  Notes: Values of 15 km and 33 GPa have been used, respectively, for the seismogenic thickness and rigidity modulus for GM1 computation. See the main text on adopted formulation. Seismic coupling coefficient (SCC) expressed as percentage of the seismic/geodetic moment-rate ratio is also reported. Uncertainties (67 per cent confidence interval) of estimated moment-rates and SCC are also reported; confidence intervals are strongly asymmetric, with large upper bounds, due to the asymmetry of the moment-rate lognormal distribution (2). View Large Estimated geodetic moment-rates (Table 1 and Fig. 5a) range in the interval 3.4 × 1017-2.2 × 1018 N·m yr−1. The higher values (≥1.5 × 1018 N·m yr−1) are observed in Southern Zagros (along the Fars arc and close to the Oman line; e.g. Z026, Z031, Z232 and Z033) and close to the Urmia and Van lakes region (Turkish–Iranian Plateau; e.g. Z001 and Z002). Seismic moment-rates (Table 1 and Fig. 5b), estimated by adopting the truncated Gutenberg–Richter relation, range in the interval 5.6 × 1015-1.3 × 1018 N·m yr−1. The higher values of the seismic moment-rates (≥1 × 1018 N·m yr−1) are observed only along the Turkish–Iranian Plateau (Z001). The Turkish–Iranian Plateau as well as Northern Zagros and the Sanandaj–Sirjan zone are also characterized by some cells with values ranging in the 5 × 1017-1 × 1018 N·m yr−1 interval (Z002, Z006, Z007, Z011, Z014 and Z019; Fig. 5b). Southern Zagros are prevailing characterized by cells with values ranging in the 2 × 1017-5 × 1017 N·m yr−1 interval, while the lower values (<1.× 1017 N·m yr−1) are observed on some cells located along the Sanandaj–Sirjan zone (Z016, Z022 and Z027; Fig. 5b). Seismic moment-rates, computed by adopting the Kostrov summation method, range in the interval 1.9 × 1015-1.3 × 1018 N·m yr−1 (Table 1 and Fig. 5b). The higher values are observed in only two cells, located nearby the Urmia lake (Z002 with a cumulative moment-rate of 9.0 × 1017 N·m yr−1) and the Buyin Zara (Z011 with a cumulative moment-rate of 1.3 × 1018 N·m yr−1) regions, respectively, which have been both characterized by the occurrence of destructive earthquakes (M > 7) during the last century (see Fig. 4a). Some cells located nearby the Van lake (Z001), the Buyin Zara zone (Z007) and on Southern Zagros (Z025, Z026, Z028, Z032 and Z033; Fig. 5b) are characterized by moment-rates ranging in the 1.0 × 1017–3.1.× 1017 N·m yr−1 interval. In the remaining sectors of the investigated area, the seismic moment-rates show values lower than 1.0 × 1017 N·m yr−1. Considering that the geodetic moment-rate is a measure of both elastic and anelastic loading rates, while the seismic moment-rate is a measure of the elastic unloading rate, the simple seismic/geodetic moment-rate ratio, expressed as a percentage, can be termed ‘seismic coupling coefficient’ (hereinafter SCC). The higher this ratio, the larger part of the measured deformation has been seismically released. Conversely, a low ratio indicates an apparent seismic moment deficit, which suggests either a proportion of aseismic deformation (i.e. ongoing unloading by creep and other plastic process) or overdue earthquakes (i.e. elastic storage). Fig. 6(a) shows the distribution of the SCC over the investigated area (estimated values are reported in Table 1; see also Fig. S9, Supporting Information). On the Turkish–Iranian Plateau and along the Sanandaj–Sirjan zone, the estimated SCC values range in the intervals 22–89 and 1–91 per cent, respectively. On Northern and Southern Zagros, the SCC values range in the intervals 20–81 and 15–33 per cent, respectively. On Fig. 6(b), we reported the SCC values resulting from the seismic/geodetic moment-rate ratio with seismic moment-rates estimated by adopting the Kostrov approach. The SCC value exceeds 50 per cent only on the two cells (Z002 and Z011), which, as previously described, have been affected by the occurrence of destructive earthquake (M > 7) during the last century. On the other cells, the SCC pattern does not exceed the value of 19 per cent (∼8 per cent on average). Figure 6. View largeDownload slide Seismic coupling coefficient (SCC), expressed as percentage of the seismic/geodetic moment-rate ratio as computed in this study (see Table 1 for details). (a) SCC from seismic moment-rates derived by the truncated Gutenberg–Richter distribution. The dotted red line marks the boundary of the evaporitic Hormuz Formation (e.g. Bahroudi & Koyi 2003). (b) SCC from seismic moment-rates derived by the cumulative Kostrov summation method. The maps are plotted in an oblique Mercator projection. Figure 6. View largeDownload slide Seismic coupling coefficient (SCC), expressed as percentage of the seismic/geodetic moment-rate ratio as computed in this study (see Table 1 for details). (a) SCC from seismic moment-rates derived by the truncated Gutenberg–Richter distribution. The dotted red line marks the boundary of the evaporitic Hormuz Formation (e.g. Bahroudi & Koyi 2003). (b) SCC from seismic moment-rates derived by the cumulative Kostrov summation method. The maps are plotted in an oblique Mercator projection. 5 DISCUSSION AND CONCLUSIONS An extensive combination of novel observations rigorously integrated with already published results improves the picture of the ongoing crustal deformation field for the entire ZFTB. Taking into account the instrumental seismicity catalogues, we have been able to provide a statistical evaluation of the seismic/geodetic deformation-rate ratio (i.e. SCC) for the area. Our GPS-based velocity field, referred to the Arabian plate, allows recognizing a clockwise rotation, passing from a ∼11–13 mm yr−1 SE-ward motion along the Turkish–Iranian Plateau to a ∼10–12 mm yr−1 SW-directed motion on the inner side of Fars arc. Sites located along the external margin of ZFTB show ∼1–3 mm yr−1 of motion, highlighting how much of the oblique Arabia–Eurasia convergence is currently absorbed within the Zagros with prevailing contractional features across Southern Zagros and right-lateral shear along the NW trending MRF in NW Zagros. Due to convergence rate variations along the general NW–SE strike of Zagros, the oblique collision is transformed by some roughly N-S-trending strike-slip faults. The dense coverage of our GPS solution allowed us to estimate ∼2.8 mm yr−1 of right-lateral motion for the Izeh–Hendijan fault and ∼3.4 mm yr−1 of right-lateral motion for the Kazerun fault (Figs 2a and 3a). No geological slip-rate estimations are currently available for the Izeh–Hendijan fault, while our estimation for the Kazerun fault matches well with previous geodetic (∼3.6 mm yr−1; Tavakoli et al.2008) and geological estimations (1.5–4 mm yr−1; Authemayou et al.2009, and references therein). These faults have a strike parallel to the Eurasia–Arabia convergence direction and, while partitioning a fraction of the convergence rate, segment the Southern Zagros range along the prevailing NS direction, as already evidenced by previous studies (e.g. Authemayou et al.2009, and references therein). We also suggest that MRF accounts for ∼8–11 mm yr−1 of right-lateral slip-rate. The right-lateral strike-slip faulting along the MRF is confirmed by focal mechanism solutions of earthquakes and tectonic geomorphology (see Talebian & Jackson 2004, and references therein). Alipoor et al. (2012) evaluated a slip-rate of 1.6–3.2 mm yr−1 along the MRF by using geological, geomorphological markers and drainage patterns. Based on GPS measurements, Vernant et al. (2004) and Walpersdorf et al. (2006) estimated 3 ± 2 and 4–6 mm yr−1 of right-lateral slip-rate for the MRF, respectively. Copley & Jackson (2006) and Authemayou et al. (2009) also suggested a slip-rate of 2–5 and 3.5–12.5 mm yr−1 along the MRF and close to its northwestern termination, respectively. As above mentioned, the lack of GPS measurements close to the MRF does not allow direct measurement of the geodetic slip-rate on this fault system. However, although our estimated rate is based on a simple vectorial decomposition (by grouping GPS velocities on each side of the fault), it concurs with the upper limit of the rates determined from long-term geomorphic offsets by Authemayou et al. (2006). The map of the SCC over the investigated area, while confirms some first-order features established by previous estimations (e.g. Jackson & McKenzie 1988; Masson et al.2005; Ansari & Zamani 2014; Khodaverdian et al.2015), leads also to some further considerations. It must be noted that previous estimates have adopted different (i) data sets, (ii) approaches and (iii) subdivision of the investigated area. The above-quoted authors, used both instrumental and historical seismicity catalogues (with different temporal coverage and magnitude threshold) and usually adopted the Kostrov approach by subdividing the area into large polygons. In addition, also the used geodetic data sets show different spatial coverages, passing from a sparse geodetic network as in Masson et al. (2005) to a dense network as in Ansari & Zamani (2014) and Khodaverdian et al. (2015). Our moment-rates and CSS estimations are based on the densest geodetic data set currently available for the investigated area and a seismic catalogue (from the ISC) spanning the 1909–2016 time interval. We are aware (i) that the ISC catalogue contains a substantial number of earthquakes which have been determined from teleseismic waveforms and that, for many of them, the location is insufficiently accurate (with uncertainties of up to 50 km; Talebian & Jackson 2004) and (ii) that such a catalogue can be considered complete for M ≥ 4.4 since 1960 (Karimiparidari et al.2013). While the uncertainty related to the poorly accurate earthquake locations cannot be properly addressed, the use of the cumulative truncated Gutenberg–Richter earthquake distribution, allow us to account for the probable incompleteness of the seismic catalogue. On the Turkish–Iranian Plateau, the SCC pattern is sampled only by four cells (Fig. 6a). Cells Z001 and Z006 are characterized by SCC values >80 per cent, while cells Z002 and Z005 show values of 30 and 22 per cent (or 60 and 33 per cent by taking into account the Shen et al. (2015) strain-rate field), respectively. On the basis of the earthquakes statistics, results achieved for cells Z001, Z002 and Z005 can be considered from good to highly reliable since them are well constrained by geodetic and seismological observations, while results achieved for cell Z006 can be viewed as adequate. Overall, achieved results well testify, beside the low SCC values inferred for cell Z005 (which have experienced only three M > 6 earthquakes in the last 1200 yr), that on the Turkish–Iranian Plateau, a large fraction (>60 per cent) of the measured crustal deformation occurs seismically. On the Sanandaj–Sirjan zone, the SCC pattern is mainly sampled by seven cells (Fig. 6a). The highest value has been inferred for cell Z011 (∼91 per cent), while, with the exception of some solitary cells (Z016, Z022 and Z027) characterized by values <10 per cent, the remaining cells (Z012, Z021 and Z028) show values in the 27–38 per cent range. Results achieved for cell Z011, although well constrained by earthquakes statistics, would be lesser reliable because the low number of GPS stations and/or the short temporal observation period (4–6 yr), therefore providing an inaccurate sample of the ongoing deformation pattern. Therefore, by excluding results achieved for cell Z011, we suggest that on Sanandaj–Sirjan zone a small to moderate fraction (<40 per cent) of the measured crustal deformation occurs seismically. On some cells (e.g. Z016 and Z027), the occurrence of large earthquakes in the past (estimated magnitudes M ≥ 6) coupled with a high geodetic deformation, could indicate overdue M ≥ 6 earthquakes. Similar considerations can be done by taking into account the moment-rates computed by using the Shen et al. (2015) strain-rate estimates (Figs S8 and S9, Supporting Information). Cells located eastward the Sanandaj–Sirjan zone, are characterized by moderate SCC values (31 and 47 per cent for cells Z023 and Z029, respectively), or by SCC values >100 per cent (cell Z007). The formers show a seismic behaviour similar to the one inferred for the Sanandaj–Sirjan zone, the latter, beside the apparent imbalance between the proxies of moment-rate, clearly highlight how a high fraction of the measured crustal deformation occurs seismically. On Northern Zagros, the SCC pattern is sampled by five cells (Fig. 6a). Cells Z014 and Z019 are characterized by SCC values of 81 and 52 per cent, respectively, while the remaining cells (Z013, Z015 and Z020) show values in the 20–36 per cent range. Results achieved for cell Z019 would be lesser reliable because Mx is constrained by only a large earthquake (M = 6.5, 840 AD; earthquakes occurred during the instrumental time do not exceed M = 5.5) whose magnitude estimation is poorly constrained (Ambraseys & Jackson 1998). This aspect suggests a possible overestimation of SCC value for cell Z019. Results achieved for cell Z014 can be considered lesser reliable also because the low number of GPS stations (that are also affected by large uncertainties) which would provide an inaccurate sample of deformation field over the investigated cell. Conversely, results achieved for cells Z015 and Z020 can be considered from good to highly reliable, since them are well constrained by earthquakes statistics and geodetic data. Considering the moment-rate values computed by using the Shen et al. (2015) strain-rate estimates (Figs S8 and S9, Supporting Information), some differences can be observed. More in detail, these last estimates are generally lower than those estimated by taking into account the ‘spline in tension’ method, leading to an increase of estimated SCC values. This aspect suggests that on this area, some local fluctuations of the strain-rate field would be better captured by the Shen et al. (2015) approach, therefore providing a finer resolution of ongoing deformation. Based on these considerations, we suggest that on Northern Zagros a moderate fraction (∼49 per cent) of the measured crustal deformation occurs seismically. On Southern Zagros, the SCC pattern is mainly sampled by five cells (Fig. 6a). Such a pattern is quite homogeneous with values ranging in the 15–33 per cent interval. On the basis of the earthquakes statistics, values estimated for cell Z026 could be considered less reliable, while for the remaining cells (Z025, Z031, Z032 and Z033) achieved results can be considered as highly reliable. Similar consideration can be done by taking into account moment-rates computed by using the Shen et al. (2015) strain-rate estimates (Figs S8 and S9, Supporting Information). These results well highlight as on Southern Zagros a large fraction of the measured crustal deformation occurs aseismically. Generally, the estimated SCC values concur well with the ones proposed by Khodaverdian et al. (2015), whereas are higher than the ones reported in the previous estimations (Jackson & McKenzie 1988; Masson et al.2005; Ansari & Zamani 2014), especially for the Zagros area. In particular, for this last area, previous estimations do not exceed the value of 15 per cent (see Ansari & Zamani 2014, and references therein). Masson et al. (2005), in estimating the SCC for the whole Iranian region, used different time intervals and observed relatively constant values for the last 100, 200 and 300 yr, concluding that an interval of 200–300 yr better depicts the SCC pattern of this area. By considering our estimations using the Kostrov approach (see the Supporting Information section for additional details), we note a general reduction of the SCC values over the investigated area. This aspect suggests that seismic moment-rates estimated with the Kostrov approach could be underestimated. Indeed, the use of the Kostrov approach suffers from the possible lack of both large earthquakes (with long recurrence interval compared to the catalogue duration) and undetected small magnitude events. As previously mentioned, the use of a statistical method such as the cumulative truncated Gutenberg–Richter earthquake distribution allows to take into account the probable incompleteness of the existing catalogue. Hence, seismic moment-rates calculated with this last approach can be assumed as representative of the long-term seismic deformation over a given region (e.g. Mazzotti et al.2011). Northern and Southern Zagros are characterized by SCC values which, based on the significant differences on the strain-rate patterns above discussed, partially overlaps. However, by taking into account the pattern of estimated seismic and geodetic moment-rates, some interesting features can be observed (Fig. 6). More in detail, while estimated seismic moment-rates are quite similar for both regions, with slight larger values inferred for Northern Zagros, the pattern of geodetic moment-rates strongly differs, with the larger values (≥1.3 × 1018 N·m yr−1) detected on the Southern Zagros (Fig. S8, Supporting Information). These features highlight that in spite of a quite similar seismic release of both regions, Southern Zagros account for the larger crustal deformation, clearly evidencing the prevailing aseismic behaviour of that region. To explain such an aseismic behaviour, a number of different hypotheses have been proposed in the last decade. For instance, Barnhart & Lohman (2013) and Barnhart et al. (2013) by using InSAR data suggested that aseismic shortening is segmented by fault creep in the sedimentary cover. Nissen et al. (2011) argued that the aseismic shortening is concentrated within the basement, as implied by shallow earthquake centroid depths. Similar conclusions have been obtained also by Allen et al. (2013) by comparing earthquake locations and topography. Others suggest that faulting in the basement occurs independently of the folding of the sedimentary cover because the cover is detached from the basement along the weak evaporitic Hormuz Formation resulting in a splitting of the seismogenic layer in two. This might explain the large number of M = 5–6 earthquakes and the complete absence of any greater than M = 7 from both instrumental and historical records (Nissen et al.2014). These observations, suggest that the seismicity pattern observed on Southern Zagros would be primarily driven by buoyancy forces arising from large lateral variations in the density and structure of the lower crust/lithosphere. A large fraction of this seismicity occurs on the upper basement and lower/mid-sedimentary cover, as suggested by the largest reverse faulting earthquakes occurred in the area (e.g. 1972 April 10 and 1977 March 21; see Nissen et al.2014, and references therein). In fact, these large earthquakes, which are located at a depth of ∼10 km and are coupled with the lack of surface rupturing, probably extended across the Hormuz Formation between the basement and the sedimentary cover, because of their obvious larger source dimensions. In addition, the incompetence of the weak evaporitic Hormuz Formation allows the occurrence of large aseismic motion on subhorizontal faults and surfaces of décollement (as suggested also by other authors; e.g. Talbot & Alavi 1996; Mouthereau et al.2007; Jahani et al.2009) hence resulting in a low SCC. Unfortunately, the possibility of splitting of the seismogenic layer in two as proposed by some investigators (Nissen et al.2014) cannot be ruled out by our data. Conversely, the deformation occurring on the Turkish–Iranian Plateau is primary driven by plate stresses which lead to a clear seismic behaviour of the area, resulting into high SCC values. In conclusion, our interpretation of geodetic and seismic data presented in this study might contribute to improve the picture of the current seismotectonic pattern of the Zagros area, clearly indicating that this active area accounts for the large deformation-rates related to the oblique Arabia–Eurasia convergence. In addition, despite a number of uncertainties, mainly related to available historical and instrumental seismic information, we constrain the aseismic fraction of the total deformation-rate budget. Estimated values clearly demonstrate how the largest fraction of aseismic deformation occurs in Southern Zagros. This is corroborating the role of massive salt deposits (the evaporitic Hormuz Formation) as possible décollement in the Southern Zagros crust (e.g. Hatzfeld & Molnar 2010, Nissen et al.2014). Aseismic deformation has generally been recognized in tectonic areas characterized by low-magnitude background seismicity (see Mazzotti et al.2011 for an overview). However, considering the high seismic activity and the large deformation-rates observed in the investigated area, an improved knowledge about the rheological behaviour of the lithosphere is required. Acknowledgements We acknowledge EUREF (www.epncb.oma.be), SOPAC (http://sopac.ucsd.edu/), UNAVCO (www.unavco.org) and NGS (http://geodesy.noaa.gov) for providing free access to GPS data. We wish to thank the Editor, Prof. Duncan Agnew, and two anonymous reviewers for their very constructive comments and suggestions, which helped us to significantly improve the early version of paper. REFERENCES Alipoor R., Zaré M., Ghassemi M.R., 2012. Inception of activity and slip rate on the Main Recent Fault of Zagros Mountains, Iran, Geomorphology , 175–176, 86– 97. https://doi.org/10.1016/j.geomorph.2012.06.025 Google Scholar CrossRef Search ADS   Allen M.B., Saville C., Blanc E.J.-P., Talebian M., Nissen E., 2013. Orogenic plateau growth: expansion of the Turkish-Iranian Plateau across the Zagros fold-and-thrust belt, Tectonics , 32( 2), 171– 190. https://doi.org/10.1002/tect.20025 Google Scholar CrossRef Search ADS   Ambraseys N.N., 1989. Temporary seismic quiescence: SE Turkey, Geophys. J. Int. , 96( 2), 311– 331. https://doi.org/10.1111/j.1365-246X.1989.tb04453.x Google Scholar CrossRef Search ADS   Ambraseys N.N., Jackson J.A., 1998. Faulting associated with historical and recent earthquakes in the Eastern Mediterranean region, Geophys. J. Int. , 133( 2), 390– 406. https://doi.org/10.1046/j.1365-246X.1998.00508.x Google Scholar CrossRef Search ADS   Ambraseys N.N., Melville C., 1982. A History of Persian Earthquakes , Cambridge University Press, Cambridge. Ansari S., Zamari A., 2014. Short-term seismic crustal deformation of Iran, Ann. Geophys. , 57( 2), S0210, doi:10.4401/ag-6413. Aqrawi A.A.M., Goff J.C., Horbury A.D., Sadooni F.N., 2010. The Petroleum Geology of Iraq , Scientific Press Ltd. ArRajehi A. et al.  , 2010. Geodetic constraints on present-day motion of the Arabian Plate: implications for Red Sea and Gulf of Aden rifting, Tectonics , 29 TC3011, 10.1029/2009TC002482. https://doi.org/doi:10.1029/2009TC002482 Google Scholar CrossRef Search ADS   Authemayou C., Chardon D., Bellier O., Malekzadeh Z., Shabanian F., Abbassi M.R., 2006. Late Cenozoic partitioning of oblique plate convergence in the Zagros fold-and-thrust belt (Iran), Tectonics , 25( 3), TC3002, doi:10.1029/2005tc001860. https://doi.org/10.1029/2005TC001860 Google Scholar CrossRef Search ADS   Authemayou C. et al.  , 2009. Quaternary slip-rates of the Kazerun and the Main Recent Faults: active strike-slip partitioning in the Zagros fold-and-thrust belt, Geophys. J. Int. , 178( 1), 524– 540. https://doi.org/10.1111/j.1365-246X.2009.04191.x Google Scholar CrossRef Search ADS   Bahroudi A., Koyi H.A., 2003. Effect of spatial distribution of Hormuz salt on deformation style in the Zagros fold and thrust belt: an analogue modelling approach, J. geol. Soc. , 160( 5), 719– 733. https://doi.org/10.1144/0016-764902-135 Google Scholar CrossRef Search ADS   Barnhart W.D., Lohman R.B., 2013. Phantom earthquakes and triggered aseismic creep: vertical partitioning of strain during earthquake sequences in Iran, Geophys. Res. Lett. , 40( 5), 819– 823. https://doi.org/10.1002/grl.50201 Google Scholar CrossRef Search ADS   Barnhart W.D., Lohman R.B., Mellors R.J., 2013. Active accommodation of plate convergence in Southern Iran: Earthquake locations, triggered aseismic slip, and regional strain rates, J. geophys. Res. , 118( 10), 5699– 5711. https://doi.org/10.1002/jgrb.50380 Google Scholar CrossRef Search ADS   Berberian M., 1995. Master blind thrust faults hidden under the Zagros folds: active basement tectonics and surface morphotectonics, Tectonophysics , 241( 3–4), 193– 224. https://doi.org/10.1016/0040-1951(94)00185-C Google Scholar CrossRef Search ADS   Beutler G., Moore A.W., Mueller I.I., 2009. The international global navigation satellite systems service (IGS): development and achievements, J. Geod. , 83( 3–4), 297– 307. https://doi.org/10.1007/s00190-008-0268-z Google Scholar CrossRef Search ADS   Bird P., 2003. An updated digital model of plate boundaries, Geochem. Geophys. Geosyst. , 4( 3), 1027, doi:10.1029/2001GC000252. https://doi.org/10.1029/2001GC000252 Google Scholar CrossRef Search ADS   Bosold A., Schwarzhans W., Julapour A., Ashrazadeh A.R., Ehsani S.M., 2005. The structural geology of the High Central Zagros revisited (Iran), Petrol. Geosci. , 11( 3), 225– 238. https://doi.org/10.1144/1354-079304-646 Google Scholar CrossRef Search ADS   Clarke P.J. et al.  , 1998. Crustal strain in central Greece from repeated GPS measurements in the interval 1989–1997, Geophys. J. Int. , 135( 1), 195– 214. https://doi.org/10.1046/j.1365-246X.1998.00633.x Google Scholar CrossRef Search ADS   Colman-Sadd S., 1978. Fold development in Zagros simply folded belt, Southwest Iran, AAPG Bull. , 62, 984– 1003. Copley A., Jackson J., 2006. Active tectonics of the Turkish-Iranian Plateau, Tectonics , 25( 6), TC6006, doi:10.1029/2005TC001906. https://doi.org/10.1029/2005TC001906 Google Scholar CrossRef Search ADS   Davies R., England P., Parsons B., Billiris H., Paradissis D., Veis G., 1997. Geodetic strain of Greece in the interval 1892–1992, J. geophys. Res. , 102( B11), 24 571-24 588. https://doi.org/10.1029/97JB01644 Google Scholar CrossRef Search ADS   Djamour Y., Vernant P., Nankali H.R., Tavakoli F., 2011. NW Iran-eastern Turkey present-day kinematics: results from the Iranian permanent GPS network, Earth planet. Sci. Lett. , 307( 1–2), 27– 34. https://doi.org/10.1016/j.epsl.2011.04.029 Google Scholar CrossRef Search ADS   Dziewonski A.M., Chou T.-A., Woodhouse J.H., 1981. Determination of earthquake source parameters from waveform data for studies of global and regional seismicity, J. geophys. Res. , 86( B4), 2825– 2852. https://doi.org/10.1029/JB086iB04p02825 Google Scholar CrossRef Search ADS   Edgell H.S, 1991. Proterozoic salt basins of the Persian Gulf area and their role in hydrocarbon generation, Precambrian Res. , 54( 1), 1– 14. https://doi.org/10.1016/0301-9268(91)90065-I Google Scholar CrossRef Search ADS   Ekström G., Nettles M., Dziewonski A.M., 2012. The global CMT project 2004-2010: centroid-moment tensors for 13,017 earthquakes, Phys. Earth planet. Inter. , 200–201, 1– 9. https://doi.org/10.1016/j.pepi.2012.04.002 Google Scholar CrossRef Search ADS   Elliott J.R. et al.  , 2015. The 2013 Mw 6.2 Khaki-Shonbe (Iran) Earthquake: insights into seismic and aseismic shortening of the Zagros sedimentary cover, Earth Space Sci. , 2( 11), 435– 471. https://doi.org/10.1002/2015EA000098 Google Scholar CrossRef Search ADS   Engdahl R.E., Jackson J.A., Myers S.C., Bergman E.A., Priestley K. 2006. Relocation and assessment of seismicity in the Iran region, Geophys. J. Int. , 167( 2), 761– 778. https://doi.org/10.1111/j.1365-246X.2006.03127.x Google Scholar CrossRef Search ADS   Frohlich C., 1992. Triangle diagrams: ternary graphs to display similarity and diversity of earthquake focal mechanisms, Phys. Earth planet. Inter. , 75( 1–3), 193– 198. https://doi.org/10.1016/0031-9201(92)90130-N Google Scholar CrossRef Search ADS   Ghasemi A., Talbot C.J., 2006. A new tectonic scenario for the Sanandaj-Sirjan Zone (Iran), J. Asian Earth Sci. , 26( 6), 683– 693. https://doi.org/10.1016/j.jseaes.2005.01.003 Google Scholar CrossRef Search ADS   Goldsmith N.F., Martinell J., Demarcq G., Bohn-Havas M., Dockery D.T.III, 1994. Sr-isotopic calibration of Cenozoic bivalvia and Early Miocene migrations: Eurasian Carnivores to Africa (the Hazeva Formation, Israel) and African gazelles and proboscidia to Ipolytarnoc, Hungary, Newslett. Stratig. , 31( 3), 167– 183. https://doi.org/10.1127/nos/31/1994/167 Google Scholar CrossRef Search ADS   Guest B., Stockli D.F., Grove M., Axen G.J., Lam P.S., Hassanzadeh J., 2006. Thermal histories from the central Alborz Mountains, northern Iran: implications for the spatial and temporal distribution of deformation in northern Iran, Bull. geol. Soc. Am. , 118( 11–12), 1507– 1521. https://doi.org/10.1130/B25819.1 Google Scholar CrossRef Search ADS   Haines A.J., Holt W.E., 1993. A procedure for obtaining the complete horizontal motions within zones of distributed deformation from the inversion of strain rate data, J. geophys. Res. , 98( B7), 12 057-12 082. https://doi.org/10.1029/93JB00892 Hanks T.C., Kanamori H., 1979. A moment magnitude scale, J. geophys. Res. , 84( B5), 12057– 12082. https://doi.org/10.1029/JB084iB05p02348 Hatzfeld D., Molnar P., 2010. Comparisons of the kinematics and deep structures of the Zagros and Himalaya and of the Iranian and Tibetan plateaus and geodynamic implications, Rev. Geophys. , 48( 2), RG2005, doi:10.1029/2009RG000304. https://doi.org/10.1029/2009RG000304 Google Scholar CrossRef Search ADS   Herring T.A., King R.W., McClusky S.C., 2010. Introduction to GAMIT/GLOBK, Release 10.4. Massachusetts Institute of Technology, Cambridge, MA, 1– 48. Hessami K., Nilforoushan F., Talbot C.J., 2006. Active deformation within the Zagros Mountains deduced from GPS measurements, J. geol. Soc. , 163( 1), 143– 148. https://doi.org/10.1144/0016-764905-031 Google Scholar CrossRef Search ADS   Hyndman R.D., Weichert D.H., 1983. Seismicity and rates of relative motion on the plate boundaries of Western North America, Geophys. J. Int. , 72( 1), 59– 82. https://doi.org/10.1111/j.1365-246X.1983.tb02804.x Google Scholar CrossRef Search ADS   Jackson J., 1992. Partitioning of strike-slip and convergent motion between Eurasia and Arabia in eastern Turkey and the Caucasus, J. geophys. Res. , 97( B9), 12 471– 12 479. https://doi.org/10.1029/92JB00944 Google Scholar CrossRef Search ADS   Jackson J.A., McKenzie D.P., 1988. The relationship between plate motions and seismic moment tensors, and the rates of active deformation in the Mediterranean and Middle East, Geophys. J. Int. , 93( 1), 45– 73. https://doi.org/10.1111/j.1365-246X.1988.tb01387.x Google Scholar CrossRef Search ADS   Jahani S., Callot J.P., Letouzey J., Frizon De Lamotte D., 2009. The eastern termination of the Zagros fold-and-thrust belt, Iran: structures, evolution, and relationships between salt plugs, folding, and faulting, Tectonics , 28( 6), doi:10.1029/2008TC002418. https://doi.org/10.1029/2008TC002418 Karimiparidari S., Zaré M., Memarian H., Kijko A., 2013. Iranian earthquakes, a uniform catalog with moment magnitudes, J. Seismol. , 17( 3), 897– 911. https://doi.org/10.1007/s10950-013-9360-9 Google Scholar CrossRef Search ADS   Kennett B.L.N., Engdahl E.R., Buland R., 1995. Constraints on seismic velocities in the Earth from traveltimes. Geophys. J. Int. , 122( 1), 108– 124. https://doi.org/10.1111/j.1365-246X.1995.tb03540.x Google Scholar CrossRef Search ADS   Kent P.E., 1979. The emergent Hormuz salt plugs of southern Iran, J Petrol. Geol. , 2( 2), 117– 144. https://doi.org/10.1111/j.1747-5457.1979.tb00698.x Google Scholar CrossRef Search ADS   Khodaverdian A., Zafarani H., Rahimian M., 2015. Long term fault slip rates, distributed deformation rates and forecast of seismicity in the Iranian Plateau, Tectonics , 34( 10), 2190– 2220. https://doi.org/10.1002/2014TC003796 Google Scholar CrossRef Search ADS   Kijko A., Graham G., 1998. Parametric-historic procedure for probabilistic seismic hazard analysis Part I: estimation ofmaximum regionalmagnitude mmax, Pure appl. Geophys. , 154( 3), 413– 442. Google Scholar CrossRef Search ADS   Kijko A., Singh M., 2011. Statistical tools for maximum possible earthquake magnitude estimation, Acta Geophys. , 59( 4), 674– 700. https://doi.org/10.2478/s11600-011-0012-6 Google Scholar CrossRef Search ADS   Kostrov B., 1974. Seismic moment and energy of earthquakes and seismic flow of rock, Izv. Acad. Sci. USSR Phys. Solid Earth , 1, 23– 40. Masson F., Anvari M., Djamour Y., Walpersdorf A., Tavakoli F., Daignieres M., Nankali H., Van Gorp S., 2007. Large-scale velocity field and strain tensor in Iran inferred from GPS measurements: new insight for the present-day deformation pattern within NE Iran, Geophys. J. Int. , 170( 1), 436– 440. https://doi.org/10.1111/j.1365-246X.2007.03477.x Google Scholar CrossRef Search ADS   Masson F., Chéry J., Hatzfeld D., Martinod J., Vernant P., Tavakoli F., Ghafory-Ashtiani M., 2005. Seismic versus aseismic deformation in Iran inferred from earthquakes and geodetic data, Geophys. J. Int. , 160( 1), 217– 226. https://doi.org/10.1111/j.1365-246X.2004.02465.x Google Scholar CrossRef Search ADS   Mazhari S.A., Bea F., Amini S., Ghalamghash J., Molina J.F., Montero P., Scarrow J.H., Williams I.S., 2009. The Eocene bimodal Piranshahr massif of the Sanandaj-Sirjan Zone, NW Iran: a marker of the end of the collision in the Zagros orogen, J. geol. Soc. , 166( 1), 53– 69. https://doi.org/10.1144/0016-76492008-022 Google Scholar CrossRef Search ADS   Mazzotti S., Adams J., 2005. Rates and uncertainties on seismic moment and deformation in eastern Canada, J. geophys. Res. , 110( B9), B09301, doi:10.1029/2004JB003510. https://doi.org/10.1029/2004JB003510 Google Scholar CrossRef Search ADS   Mazzotti S., Leonard L.J., Cassidy J.F., Rogers G.C., Halchuk S., 2011. Seismic hazard in western Canada from GPS strain rates versus earthquake catalog, J. geophys. Res. , 116( B12), B12310, doi:10.1029/2011JB008213. https://doi.org/10.1029/2011JB008213 Google Scholar CrossRef Search ADS   McQuarrie N. 2004. Crustal scale geometry of the Zagros fold-thrust belt, Iran, J. Struct. Geol. , 26( 3), 519– 535. https://doi.org/10.1016/j.jsg.2003.08.009 Google Scholar CrossRef Search ADS   Melville C., 1984. The use of Historical records for seismic assessment, Boll. Geofis. Teor. Appl. , 26, 109– 119. Mirzaei N., Gao M., Chen Y.T., Wang J., 1997. A uniform catalog of earthquakes for seismic hazard assessment in Iran, Acta Seismol. Sin. , 10( 6), 713– 726. https://doi.org/10.1007/s11589-997-0003-5 Google Scholar CrossRef Search ADS   Molinaro M., Zeyen H., Laurencin X., 2005. Lithospheric structure beneath the south-eastern Zagros Mountains, Iran: recent slab break-off?, Terra Nova , 17( 1), 1– 6. https://doi.org/10.1111/j.1365-3121.2004.00575.x Google Scholar CrossRef Search ADS   Mousavi, Z., Walpersdorf A., Walker R.T., Tavakoli F., Pathier E., Nankali H., Nilfouroushan F., Djamour Y., 2013. Global Positioning System constraints on the active tectonics of NE Iran and the South Caspian region, Earth planet. Sci. Lett. , 377–378, 287– 298. https://doi.org/10.1016/j.epsl.2013.07.007 Google Scholar CrossRef Search ADS   Mouthereau F., Lacombe O., Vergés J., 2012. Building the Zagros collisional orogen: timing, strain distribution and the dynamics of Arabia/Eurasia plate convergence, Tectonophysics , 532, 27– 60. https://doi.org/10.1016/j.tecto.2012.01.022 Google Scholar CrossRef Search ADS   Mouthereau F., Tensi J., Bellahsen N., Lacombe O., De Boisgrollier T., Kargar S., 2007. Tertiary sequence of deformation in a thin-skinned/thick-skinned collision belt: the Zagros folded belt (Fars, Iran), Tectonics , 26( 5), Tc5006, doi:10.1029/2007tc002098. https://doi.org/10.1029/2007TC002098 Google Scholar CrossRef Search ADS   Nissen E., Jackson J., Jahani S., Tatar M., 2014. Zagros “phantom earthquakes” reassessed-The interplay of seismicity and deep salt flow in the Simply Folded Belt?, J. geophys. Res. , 119( 4), doi:10.1002/2013JB010796. https://doi.org/10.1002/2013JB010796 Nissen E., Tatar M., Jackson J.A., Allen M.B., 2011. New views on earthquake faulting in the Zagros fold-and-thrust belt of Iran, Geophys. J. Int. , 186( 3), 928– 944. https://doi.org/10.1111/j.1365-246X.2011.05119.x Google Scholar CrossRef Search ADS   Okay A.I., Zattin M., Cavazza W., 2010. Apatite fission-track data for the Miocene Arabia-Eurasia collision, Geology , 38( 1), 35– 38. https://doi.org/10.1130/G30234.1 Google Scholar CrossRef Search ADS   Palano M., 2015. On the present-day crustal stress, strain-rate fields and mantle anisotropy pattern of Italy, Geophys. J. Int. , 200( 2), 969– 985. https://doi.org/10.1093/gji/ggu451 Google Scholar CrossRef Search ADS   Palano M., Imprescia P., Gresta S., 2013. Current stress and strain-rate fields across the Dead Sea Fault System: constraints from seismological data and GPS observations, Earth planet. Sci. Lett. , 369–370, 305– 316. https://doi.org/10.1016/j.epsl.2013.03.043 Google Scholar CrossRef Search ADS   Paul A., Kaviani A., Hatzfeld D., Vergne J., Mokhtari M., 2006. Seismological evidence for crustal-scale thrusting in the Zagros mountain belt (Iran), Geophys. J. Int. , 166( 1), 227– 237. https://doi.org/10.1111/j.1365-246X.2006.02920.x Google Scholar CrossRef Search ADS   Peyret M. et al.  , 2009. Present-day strain distribution across the Minab-Zendan-Palami fault system from dense GPS transects, Geophys. J. Int. , 179( 2), 751– 762. https://doi.org/10.1111/j.1365-246X.2009.04321.x Google Scholar CrossRef Search ADS   Reilinger R. et al.  , 2006. GPS constraints on continental deformation in the Africa-Arabia-Eurasia continental collision zone and implications for the dynamics of plate interactions, J. geophys. Res. , 111( B5), B05411, doi:10.1029/2005JB004051. https://doi.org/10.1029/2005JB004051 Google Scholar CrossRef Search ADS   Sattarzadeh Y., Cosgrove J.W., Vita-Finzi C., 1999. The interplay of faulting and folding during the evolution of the Zagros deformation belt, in Forced Folds and Fractures , Vol. 169, pp. 187– 196, eds, Cosgrove J.W., Ameen M.S., Geol. Soc. Spec. Publ. Google Scholar CrossRef Search ADS   Savage J.C., Simpson R.W., 1997. Surface strain accumulation and the seismic moment tensor, Bull. seism. Soc. Am. , 87, 1345– 1353. Shen Z.-K., Wang M., Zeng Y., Wang F., 2015. Optimal interpolation of spatially discretized geodetic data, Bull. seism. Soc. Am. , 105( 4), 2117– 2127. https://doi.org/10.1785/0120140247 Google Scholar CrossRef Search ADS   Sherkati S., Letouzey J., Frizon de Lamotte D. 2006. Central Zagros fold-thrust belt (Iran): new insights from seismic data, field observation, and sandbox modeling, Tectonics , 25( 4), TC4007, doi:10.1029/2004TC001766. https://doi.org/10.1029/2004TC001766 Google Scholar CrossRef Search ADS   Smith W.H.F., Wessel P., 1990. Gridding with continuous curvature splines in tension, Geophysics , 55( 3), 293– 305. https://doi.org/10.1190/1.1442837 Google Scholar CrossRef Search ADS   Talbot C.J., Alavi M., 1996. The past of a future syntaxis across the Zagros, in Salt Tectonics , Vol. 100, pp. 89– 109, eds, Alsop G.I., Blundell D.J., Davison I., Geol. Soc. Spec. Publ.. Google Scholar CrossRef Search ADS   Talebian M., Jackson J., 2004. A reappraisal of earthquake focal mechanisms and active shortening in the Zagros mountains of Iran, Geophys. J. Int. , 156( 3), 506– 526. https://doi.org/10.1111/j.1365-246X.2004.02092.x Google Scholar CrossRef Search ADS   Tavakoli F. et al.  , 2008. Distribution of the right-lateral strike-slip motion from the Main Recent Fault to the Kazerun Fault System (Zagros, Iran): evidence from present-day GPS velocities, Earth planet. Sci. Lett. , 275( 3–4), 342– 347. https://doi.org/10.1016/j.epsl.2008.08.030 Google Scholar CrossRef Search ADS   Tchernov E., Ginsburg L., Tassy P., Goldsmith N.F., 1987. Miocene mammals of the Negev (Israel), J. Verteb. Paleontol. , 7( 3), 284– 310. https://doi.org/10.1080/02724634.1987.10011661 Google Scholar CrossRef Search ADS   Vernant P., et al., 2004. Contemporary crustal deformation and plate kinematics in middle east constrained by GPS measurement in Iran and northern Oman, Geophys. J. Int. , 157, 381– 398. Google Scholar CrossRef Search ADS   Walpersdorf A., et al., 2006. Difference in the GPS deformation pattern of North and Central Zagros (Iran), Geophys. J. Int. , 167( 3), 1077– 1088. https://doi.org/10.1111/j.1365-246X.2006.03147.x Google Scholar CrossRef Search ADS   Walpersdorf A. et al.  , 2014. Present-day kinematics and fault slip rates in Eastern Iran, derived from 11 years of GPS data, J. geophys. Res. , 119( 2), 1359– 1383. https://doi.org/10.1002/2013JB010620 Google Scholar CrossRef Search ADS   Weichert D.H., 1980. Estimation of the earthquake recurrence parameters from unequal observation periods for different magnitudes, Bull. seism. Soc. Am. , 70, 1337– 1346. Wells D.L., Coppersmith K.J., 1994. New empirical relationship among magnitude, rupture length, rupture width, rupture area, and surface displacement, Bull. seism. Soc. Am. , 84, 974– 1002. Wessel P., Bercovici D., 1998. Interpolation with splines in tension: a Green's function approach, Math. Geol. , 30( 1), 77– 93. https://doi.org/10.1023/A:1021713421882 Google Scholar CrossRef Search ADS   Zarifi Z., Nilfouroushan F., Raeesi M., 2014. Crustal Stress Map of Iran: insight from seismic and geodetic computations, Pure appl. Geophys. , 171( 7), 1219– 1236. https://doi.org/10.1007/s00024-013-0711-9 Google Scholar CrossRef Search ADS   Zoback M.L., 1992. First- and second-order patterns of stress in the lithosphere: the World Stress Map Project, J. geophys. Res. , 97( B8), 11 703– 11 728. https://doi.org/10.1029/92JB00132 Google Scholar CrossRef Search ADS   SUPPORTING INFORMATION Supplementary data are available at GJI online. Figure S1. (a) Strain-rate field according to a Gaussian function for distance weighting and Voronoi cell for areal weighting. (b) Strain-rate field computed with a quadratic function for distance weighting and Voronoi cell for areal weighting. For both computations, we used a weighting threshold value of 24. Strain-rates derived from the velocity interpolation only (the grid cell contain zero or just one GPS site) are reported as white arrows. Uncertainties are also reported. Figure S2. Second invariant from strain-rate fields computed by adopting (a) a ‘spline in tension’ function in deriving a continuous velocity gradient tensor over the study area, (b) a Gaussian function for distance weighting and Voronoi cell for areal weighting, (c) a quadratic function for distance weighting and Voronoi cell for areal weighting. Figure S3. Magnitude histogram plot for earthquakes collected in the database. Figure S4. Time–magnitude plot for earthquakes collected in the database. Figure S5. Depth distribution of earthquakes collected in the database. Figure S6. Examples of cumulative frequency–magnitude distributions (blue diamonds) of earthquakes for selected cells. The red line represents the truncated Gutenberg–Richter function according to eq. (2). Parameters used in eq. (2) are reported in Tables S6 and S8. Figure S7. Observed (in blue; date of the event is also reported) and estimated (in red; associated uncertainties are also reported; see and Table S8 for details) maximum magnitude over the investigated area. Cells coloured in light red have been omitted from the computation (see the main text for details). Figure S8. Moment-rate estimates computed in this study (see Table 1 in the main text for details). GM: geodetic moment-rates computed by adopting a ‘spline in tension’ approach (1) and Gaussian function (2) for distance weighting and Voronoi cell for areal weighting in deriving a continuous velocity gradient tensor over the study area. SM: seismic moment-rates computed by adopting a cumulative truncated Gutenberg–Richter distribution (GR) and a Kostrov summation method (K). Uncertainties (67 per cent confidence interval) are also reported; uncertainties for SM(K) can be considered as ∼5–10 per cent of the estimated value, reflecting a standard error of 0.2 on earthquake magnitudes. Figure S9. Comparison of the seismic coupling coefficient (expressed as a percentage) computed as: SCC(1), ratio between SM(GR) and GM(1) and SCC(2), ratio between SM(GR) and GM(2). Abbreviations are as Fig. S8. Table S1. Historical events used in this study as collected from MIRZ97 (Mirzaei et al.1997), AM&JA98 (Ambraseys & Jackson 1998) and BER95 (Berberian 1995). This table has been provided as text file. Table S2. Site coordinates and velocities (mm yr−1) referred to ITRF2008 for sites used to define the Euler parameters for the fixed Arabian plate. Uncertainties are within the 1σ confidence level. Residual velocities with respect the Arabian plate are also reported. (*) Solutions from ArRajehi et al. (2010). This table has been provided as text file. Table S3. Euler pole parameters, associated errors (3σ) and covariance matrix for the Arabian plate. Equivalent rotation vector Ω (ωx, ωy, ωz) in a Cartesian frame (and associated standard deviation) are as follow: 0.3323 ± 0.0037, −0.0194 ± 0.0036 and 0.4079 ± 0.0028 (deg Myr−1 units). Table S4. Site coordinates, velocities (mm yr−1) referred to ITRF2008 and Arabian plate used in this study. Uncertainties are within the 1σ confidence level. References are also reported. Table S5. Database of instrumental seismicity used in this study. The database has been collected from the ISC online catalogue. Table S6. Summary of earthquake catalogue statistic. NEq, number of earthquakes; Mc, Magnitude of completeness; a and b, seismicity level and slope of the Gutenberg–Richter recurrence relation and associated uncertainty; dT, time interval. This table has been provided as text file. Table S7. Summary of maximum magnitude Mx estimation by applying the 12 procedures implemented in the MMAX toolbox (Kijko & Singh 2011) and by using as input our instrumental database. We reported also the highest magnitude recorded for each cell in the instrumental (see Table S5) and/or historical catalogues (Mirzaei et al.1997; Ambraseys & Jackson 1998; see Table S1 for details). Abbreviations: N-P-G, Non-Parametric with Gaussian kernel; N-P-OS, Non-Parametric procedure based on order statistic; LM, procedure based on few a largest earthquake, nL1, L1 norm regression; LS, Least square approach; T-P, Tate−Pisarenko procedure; K-S_a, Kijko–Sellevol with Cramér's approximation; K-S_e, Kijko–Sellevol (exact solution); T-P-B, Tate–Pisarenko–Bayes procedure; K-S-B, Kijko–Sellevoll–Bayes procedure, R-W, Robson–Whitock, R-W-C, Robson–Whitock–Cooke. Some cells have been omitted because the low number of earthquakes (Z008 and Z017; see Table S6) or because not used for the moment-rate comparisons (Z003, Z004, Z009, Z010, Z018, Z024 and Z030). Table S8. Summary of maximum magnitude Mx estimation by applying the 12 procedures implemented in the MMAX toolbox (Kijko & Singh 2011) and by taking into account a catalogue which include both instrumental (see Table S5) and historical seismic records (see Table S1). We reported also the highest magnitude recorded for each cell in the instrumental (Table S5) and/or historical catalogues (Mirzaei et al.1997; Ambraseys & Jackson 1998; Table S1). Abbreviations are as Table S7. Among the different estimations, for each cell, the Mx value having the smallest uncertainties (cells coloured in grey) have been chosen for the seismic moment-rate estimation. Some cells have been omitted because the low number of earthquakes (Z008 and Z017; see Table S6) or because not used for the moment-rate comparisons (Z003, Z004, Z009, Z010, Z018, Z024 and Z030). Please note: Oxford University Press is not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the paper. © The Author(s) 2017. Published by Oxford University Press on behalf of The Royal Astronomical Society. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Geophysical Journal International Oxford University Press

# An improved evaluation of the seismic/geodetic deformation-rate ratio for the Zagros Fold-and-Thrust collisional belt

, Volume 213 (1) – Apr 1, 2018
16 pages

/lp/ou_press/an-improved-evaluation-of-the-seismic-geodetic-deformation-rate-ratio-Lf1uDGb1wL
Publisher
The Royal Astronomical Society
ISSN
0956-540X
eISSN
1365-246X
D.O.I.
10.1093/gji/ggx524
Publisher site
See Article on Publisher Site

### Abstract

Summary We present an improved picture of the ongoing crustal deformation field for the Zagros Fold-and-Thrust Belt continental collision zone by using an extensive combination of both novel and published GPS observations. The main results define the significant amount of oblique Arabia–Eurasia convergence currently being absorbed within the Zagros: right-lateral shear along the NW trending Main Recent fault in NW Zagros and accommodated between fold-and-thrust structures and NS right-lateral strike-slip faults on Southern Zagros. In addition, taking into account the 1909–2016 instrumental seismic catalogue, we provide a statistical evaluation of the seismic/geodetic deformation-rate ratio for the area. On Northern Zagros and on the Turkish–Iranian Plateau, a moderate to large fraction (∼49 and >60 per cent, respectively) of the crustal deformation occurs seismically. On the Sanandaj–Sirjan zone, the seismic/geodetic deformation-rate ratio suggests that a small to moderate fraction (<40 per cent) of crustal deformation occurs seismically; locally, the occurrence of large historic earthquakes (M ≥ 6) coupled with the high geodetic deformation, could indicate overdue M ≥ 6 earthquakes. On Southern Zagros, aseismic strain dominates crustal deformation (the ratio ranges in the 15–33 per cent interval). Such aseismic deformation is probably related to the presence of the weak evaporitic Hormuz Formation which allows the occurrence of large aseismic motion on both subhorizontal faults and surfaces of décollement. These results, framed into the seismotectonic framework of the investigated region, confirm that the fold-and-thrust-dominated deformation is driven by buoyancy forces; by contrast, the shear-dominated deformation is primary driven by plate stresses. Creep and deformation, Satellite geodesy, Asia, Earthquake hazards, Seismicity and tectonics 1 INTRODUCTION Geodetic and seismic strain-rate comparison provides significant insights into the seismic hazard of regions subject to relevant tectonic deformation. In particular, when the coseismic displacement is lower than the geodetic displacement across faults, the excess geodetic strain can be released either through large impending earthquakes or in aseismic mode (e.g. Davies et al.1997; Clarke et al.1998). Although this approach has been applied to several regions worldwide, there are several uncertainties regarding the physical significance of the deformation-rates mismatch over varying spatial and temporal scales. Roughly speaking, geodetically observed strains may include both elastic and anelastic components, and in many cases, it is difficult to differentiate the two without a priori knowledge of the rheology of the crust under investigation. Because only the elastic strain is released by earthquakes, the comparison of geodetic and seismic strain-rates may not balance in regions cut by creeping faults or where significant deformation takes place plastically. Thus, verifying actual mismatch between the geodetic and the seismic deformation-rate and estimating its extent require a high level of accuracy and spatiotemporally dense geodetic observations. Apparent imbalance between these proxies of strain-rate may also occur where the seismic catalogue is incomplete, as in the case where the seismic cycle is longer than the duration of the observation period. The Zagros Fold-and-Thrust Belt (ZFTB) continental collision zone is one of the youngest and most seismically active zones on Earth (Fig. 1). By using both instrumental and historical seismicity data coupled with deformation-rates deduced from plate movements, Jackson & McKenzie (1988) made the first quantitative comparison between seismic and deformation-rates of ZFTB. Based on a combined analysis of geodetic and seismic data that had accumulated over the years, Masson et al. (2005) provided a more recent estimation of the contemporaneous seismic versus geodetic deformation-rate ratio. They observed that a low seismic strain-rate, especially in the southern sector of ZFTB, accounted for a small percentage of the observed deformation-rate. This has indicated that the crust is deforming mostly aseismically, or that elastic strain is being accumulated. Figure 1. View largeDownload slide Regional plate tectonic setting of the study region and surrounding areas. Points and hexagons indicate GPS sites used in this work: GPS sites processed in this study are coloured in yellow, while GPS data published by other authors are reported as black symbols (see Table S4, Supporting Information). GPS sites used to estimate the Euler vector for the Arabian plate are reported as hexagons (see the Supporting Information section for details). Plate boundaries are marked as red lines (Bird 2003). Abbreviations are as follows: BS, Black Sea; CS, Caspian Sea; EAF, Eastern Anatolian fault; MS, Mediterranean Sea; SB, Sinai Block; PG, Persian Gulf; GO, Gulf of Oman; AS, Arabian Sea and Me, Mesopotamia. Dashed box shows the study area. The map is plotted in a Mercator projection. Figure 1. View largeDownload slide Regional plate tectonic setting of the study region and surrounding areas. Points and hexagons indicate GPS sites used in this work: GPS sites processed in this study are coloured in yellow, while GPS data published by other authors are reported as black symbols (see Table S4, Supporting Information). GPS sites used to estimate the Euler vector for the Arabian plate are reported as hexagons (see the Supporting Information section for details). Plate boundaries are marked as red lines (Bird 2003). Abbreviations are as follows: BS, Black Sea; CS, Caspian Sea; EAF, Eastern Anatolian fault; MS, Mediterranean Sea; SB, Sinai Block; PG, Persian Gulf; GO, Gulf of Oman; AS, Arabian Sea and Me, Mesopotamia. Dashed box shows the study area. The map is plotted in a Mercator projection. Here, we present an improved picture of the ongoing crustal deformation field of ZFTB, based on an extensive combination of novel observations rigorously integrated with published geodetic velocities. In addition, we compare the GPS (Global Positioning System)-based moment-rates with those derived from earthquake catalogue in order to provide an updated statistical evaluation of the seismic/geodetic deformation-rate ratio for the whole ZFTB. Previous studies, in estimating the seismic/geodetic deformation-rate ratio over the ZFTB, simply used the existing (historical and instrumental) record of seismicity; here we define the long-term moment release rate by adopting a truncated Gutenberg–Richter relation. This method has the advantage of being insensitive to the short duration of the observation period compared with the typical length of the seismic cycle. 2 BACKGROUND SETTING 2.1 Geological and tectonic setting The ZFTB represents the present-day boundary between the Arabian and Eurasian plates (Fig. 1). ZFTB extends for about 2000 km from eastern Turkey where it connects with the Eastern Anatolian Fault (EAF), through SW Iran to the Oman Sea, where it connects to the Makran subduction zone. The belt varies in width (from ∼150 km in the west to ∼300 km in the east) with two main embayments (Kirkuk and Dezful) and two main arcs (Lorestan and Fars; Fig. 2). Figure 2. View largeDownload slide (a) Simplified structural map of the Zagros area derived from Berberian (1995) and Allen et al. (2013), and references therein. A number of faults are inferred from changes of stratigraphic level, and are not directly observed in the field (Berberian 1995). The dashed blue line is a simplified representation of MRF. The 2° × 2° grid (and associated id) used for the geodetic and seismic moment-rate computations is reported as blue lines. Abbreviations are as follows: MRF, Main Recent Fault; MZRF, Main Zagros Reverse Fault; HZF, High Zagros Fault; KhF, Khanaqin Fault, IHF, Izeh–Hendijan Fault; KaF, Kazerun Fault. KaF divides the Zagros into Northern and Southern Zagros. (b) Historical earthquakes (yellow squares) occurring during 800–1908 CE are also reported (Mirzaei et al.1997; Ambraseys & Jackson 1998; see also Table S1 in Supporting Information). The map is plotted in an oblique Mercator projection. Figure 2. View largeDownload slide (a) Simplified structural map of the Zagros area derived from Berberian (1995) and Allen et al. (2013), and references therein. A number of faults are inferred from changes of stratigraphic level, and are not directly observed in the field (Berberian 1995). The dashed blue line is a simplified representation of MRF. The 2° × 2° grid (and associated id) used for the geodetic and seismic moment-rate computations is reported as blue lines. Abbreviations are as follows: MRF, Main Recent Fault; MZRF, Main Zagros Reverse Fault; HZF, High Zagros Fault; KhF, Khanaqin Fault, IHF, Izeh–Hendijan Fault; KaF, Kazerun Fault. KaF divides the Zagros into Northern and Southern Zagros. (b) Historical earthquakes (yellow squares) occurring during 800–1908 CE are also reported (Mirzaei et al.1997; Ambraseys & Jackson 1998; see also Table S1 in Supporting Information). The map is plotted in an oblique Mercator projection. The Main Zagros Reverse Fault (MZRF), juxtaposing the Palaeozoic-Cretaceous stratigraphy against the low-grade metamorphics and Cretaceous cover sequence of the Sanandaj–Sirjan metamorphic belt, is considered as the suture between the Arabian plate and the Iranian Plateau and represents the NE limit of the ZFTB (e.g. Authemayou et al.2006; Paul et al.2006, and references therein). The Main Recent fault (MRF), a young seismically active right-lateral strike-slip fault with a prevailing NW–SE strike, interconnects the MZRF with the EAF termination (Copley & Jackson 2006; Figs 1 and 2). Southwest of the MZRF, the Zagros Mountains are usually divided into the High Zagros (or Imbricate Belt) to the NE and the Simply Folded Zone to the SW. The former, exposing stratigraphic levels in the Mesozoic and Palaeozoic ages, lies between the original suture (i.e. MZRF) and a major thrust, the High Zagros Fault (HZF), which is mapped as running roughly parallel to the suture (Berberian 1995; Bosold et al.2005). The latter, exposing a Mesozoic–Cenozoic (Neotethys) mixed carbonate-clastic succession (Palaeozoic strata are rarely exposed), lies between the HZF, the Persian Gulf and the neighbouring plains of Mesopotamia representing the current flexural foredeep of the ZFTB (e.g. Aqrawi et al.2010, and references therein). In the Fars, these cover sequences are decoupled from the underlying basement along the Lower Cambrian evaporitic Hormuz Formation (1–3 km thick), which has played a significant role in controlling the deformation of the cover (McQuarrie 2004; Molinaro et al.2004; Jahani et al.2009). This formation has emerged as numerous salt diapirs at the surface, and has brought fragments of igneous and metamorphic rocks from the underlying basement (Kent 1979). The evaporitic Hormuz Formation is present across a wide area of the Zagros and Middle East (e.g. Edgell 1991) and contains thick evaporites, mainly halite. However, the lack of diapirs and plugs in the Dezful embayment and Lorestan arc has been attributed to the absence of the Hormuz Formation in these areas (Talbot & Alavi 1996). Moreover, the presence of a number of post-Triassic evaporitic or shale layers has been recognized in different regions of the belt and has been interpreted as possible intermediate décollement levels within the sedimentary cover (Sherkati et al.2006). The ZFTB involves parallel folds that have formed by buckling above the evaporitic Hormuz Formation (Colman-Sadd 1978; McQuarrie 2004; Mouthereau et al.2007), thrust-cored anticlines (Sattarzadeh et al.1999; Molinaro et al.2004), and thrusts verging both to the SW and to the NE (the last category responsible for some recent large earthquakes; e.g. Nissen et al.2014; Elliott et al.2015) in the sedimentary cover and its underlying basement, implying both thin- and thick-skinned shortening across the belt (Talbot & Alavi 1996; Mouthereau et al.2007). The N-S Arabia–Eurasia convergence process seems transformed by some roughly N-S-trending strike-slip faults (e.g. the Izeh−Hendijan, Kazerun and Khanaqin right-lateral faults; Fig. 2a). Among these faults, the Kazerun fault constitutes the western limit of the evaporitic Hormuz basin (Edgell 1991). The timing of initial Arabia–Eurasia collision is still debated, with recently published estimates spanning from the Late Cretaceous/Palaeocene (Ghasemi & Talbot 2006; Mazhari et al.2009), the Early Miocene (Okay et al.2010) to the Middle/Late Miocene (Guest et al.2006). Moreover, opinions are now converging to a Latest Eocene (∼35 Ma) age for initial collision (when the distal Arabian continental margin, driven by its negative buoyancy, was underthrusted beneath the upper Iranian block) and to a Late Oligocene (∼25 Ma) age for the onset of the continental crustal thickening (see Mouthereau et al.2012, and references therein for a review). Palaeozoological evidence suggests that the Arabian plate land bridge for large mammal migration between India and Africa initiated only as late as 17 Ma (e.g. Tchernov et al.1987; Goldsmith et al.1994). 2.2 Historical seismicity Available historical seismic catalogues for the Zagros area document the occurrence of large (M > 6) earthquakes since AD 200. However, the accuracy of these catalogues is non-uniform and questionable mainly due to the sparsely populated area (e.g. Mirzaei et al.1997; Ambraseys & Jackson 1998, and references therein). Studies focusing on the completeness of available data in the study area and surrounding regions (Ambraseys & Melville 1982; Melville 1984; Ambraseys 1989) have pointed out that many small and moderate earthquakes have most likely been overlooked and that available catalogues can be reasonably considered complete for earthquakes with M ≥ 6 since 1860 (Mirzaei et al.1997). Historic earthquakes striking the area during the last millennium are largely concentrated on Southern Zagros and along the Turkish–Iranian Plateau (Fig. 2b; Mirzaei et al.1997; Ambraseys & Jackson 1998), where the MRF connects with the EAF termination across an 80-km-wider shear zone. Such a shear zone is made by numerous subparallel NW–SE striking faults characterized by right-lateral strike-slip kinematics (e.g. Copley & Jackson 2006). In detail, the stronger historic earthquakes (M ≥ 7.0; AD 1008, AD 1042, AD 1696, AD 1721, AD 1780, AD 1840, AD 1903; see also Table S1, Supporting Information section for details) are mainly concentrated on the Turkish–Iranian Plateau, while the Southern Zagros have experienced only a single large earthquake (estimated magnitude M = 7.1) in AD 1440. 3 GEODETIC AND INSTRUMENTAL SEISMIC DATA 3.1 GPS data Here, we analysed an extensive GPS data set covering ∼15 yr of observations, from 1999.00 up to 2014.00, and covering the Arabian plate and surrounding areas. GPS data come from public online archives, which include CDDIS (https://cddis.nasa.gov), EUREF (www.epncb.oma.be), SOPAC (http://sopac.ucsd.edu/), UNAVCO (www.unavco.org) and NGS (http://geodesy.noaa.gov). Raw GPS data were processed by using the GAMIT/GLOBK software (Herring et al.2010; www-gpsg.mit.edu), by adopting the strategy described in Palano (2015). By using the GLORG module of GLOBK, all the GAMIT solutions and their full covariance matrices were combined to estimate a consistent set of positions and velocities in the ITRF2008 reference frame. Our analysed network does not cover entirely the Zagros area. Nonetheless, since our solution shares some stations with the ones processed by Hessami et al. (2006), Reilinger et al. (2006), Walpersdorf et al. (2006, 2014), Masson et al. (2007), Tavakoli et al. (2008), Peyret et al. (2009), Djamour et al. (2011), Mousavi et al. (2013), Palano et al. (2013) and Zarifi et al. (2013), we rigorously integrated data sets by applying a Helmert transformation of the different estimated velocity fields, in order to generate a combined velocity field aligned to a unified frame such as the ITRF2008 one. To adequately show the crustal deformation pattern over the area investigated, we rotate the unified ITRF2008 GPS velocity solution into an Arabian fixed reference frame (Fig. 3a; see Tables S2–S4, Supporting Information section for additional details). Figure 3. View largeDownload slide (a) GPS velocities and 95 per cent confidence ellipses in a fixed Arabian plate (see Table S2 and S3, Supporting Information section for details). Abbreviations are as Fig. 2. The dashed blue line is a simplified representation of MRF. The dashed dark-blue lines represent the expected Arabia-Eurasia relative motion direction. (b) Geodetic strain-rate field and associate uncertainties: arrows represent the greatest extensional (red) and contractional (blue) horizontal strain-rates. Strain-rates derived from the velocity interpolation only (the grid cell contain zero or just one GPS site) are reported as white colour. The blue lines defined the 2° × 2° grid used for the geodetic and seismic moment-rates computation. Yellow dots stand for GPS sites. The map is plotted in an oblique Mercator projection. Figure 3. View largeDownload slide (a) GPS velocities and 95 per cent confidence ellipses in a fixed Arabian plate (see Table S2 and S3, Supporting Information section for details). Abbreviations are as Fig. 2. The dashed blue line is a simplified representation of MRF. The dashed dark-blue lines represent the expected Arabia-Eurasia relative motion direction. (b) Geodetic strain-rate field and associate uncertainties: arrows represent the greatest extensional (red) and contractional (blue) horizontal strain-rates. Strain-rates derived from the velocity interpolation only (the grid cell contain zero or just one GPS site) are reported as white colour. The blue lines defined the 2° × 2° grid used for the geodetic and seismic moment-rates computation. Yellow dots stand for GPS sites. The map is plotted in an oblique Mercator projection. In addition, by taking into account the observed horizontal velocity field and associated uncertainties, we derived a continuous velocity gradient over the study area on a regular 1° × 1° grid (whose nodes do not coincide with any of the GPS stations) using a ‘spline in tension’ function (Wessel & Bercovici 1998). The 1° value corresponds to the average distance between GPS stations (in degrees). Sites biased by large velocity uncertainties and/or showing suspicious movements with respect to nearby sites (∼1 per cent of the data set here analysed) were not used for the velocity gradient computation. The tension is controlled by a factor T, where T = 0 leads to a minimum curvature (natural bicubic spline), while T = 1 leads to a maximum curvature, allowing for maxima and minima only at observation points (Smith & Wessel 1990). We set T to the value of 0.5, because it represents the optimal value to minimize short wavelength noise (see Palano 2015, and references therein). Lastly, we computed the average 2-D strain-rate tensor (and its standard error) as derivative of the velocities at the centre of each cell. The estimated strain-rates are shown in Fig. 3(b) as principal extensional ($${\dot{\varepsilon }_{H\max }}$$) and shortening ($${\dot{\varepsilon }_{h\min }}$$) horizontal strain-rates. Various methods can be used to derive GPS strain-rates, ranging from simple Delaunay triangulations to more complex parametric inversions (e.g. Haines & Holt 1993), with significant variations in their results. Therefore, since the spatial distribution of our velocity field data is heterogeneous, in order to assess the first-order variability in our strain-rate analysis and the derived moment-rates, we undertook some additional computations (see the Supporting Information section) by adopting the method described in Shen et al. (2015). This method, in interpolating the displacement field, introduces the spatial weighting function of data in various forms (e.g. uniform Gaussian or quadratic spatial weighting function), allowing to obtain finer resolution especially for regions in which data are sparsely distributed. It must be noted that some of the cells contain nil or just a single GPS site and the strain-rates information derived from the velocity interpolation only; because these estimates cannot be considered accurate for the area due to the lack of spatial constraints from data, in the following, information coming from these cells have been omitted. Overall, the GPS-based velocity field (referred to an Arabian fixed reference frame) clearly depicts a clockwise rotation, passing from an SE-ward motion along the Turkish–Iranian Plateau to an SW-directed motion on the Fars Arc. This rotation implies that in the former, the velocity field shows an oblique relationship (more than 25°) in comparison with the predicted convergent motion of Arabia with respect to Eurasia plate, while in the latter the velocity field is near parallel to such a predicted motion (Fig. 3a). Because the trend of the Zagros does not change along strike, both the above-mentioned features imply that shortening across the northwestern sector contributes less to that overall Eurasia–Arabia convergence than does convergence across the southeastern sector, as evidenced by previous studies (e.g. Jackson 1992; Copley & Jackson 2006; Reilinger et al.2006). Descriptions refining these main features are detailed below. Stations located along the Turkish–Iranian Plateau and along the NE border of ZFTB show a small clockwise rotation passing from an SE-directed motion near the Urmia Lake region (Fig. 3a rates of ∼11–12 mm yr−1) to an SSE-directed motion (Fig. 3a rates of ∼9–11 mm yr−1) close to the Qom region. Conversely, stations located in the region spanning the Kirkuk embayment, the Lorestan arc, the MRF (Fig. 3a) and westward the ZFTB frontal area, in the stable Arabian plate (e.g. ISBA, ISSD and ISNA stations), show negligible deformation. Considering the different motion occurring across MRF, a general right-lateral shear of this area, can be deduced. The lack of GPS measurements close to the MRF does not allow direct measurement of the geodetic slip-rate on this fault system; however, considering the pattern previously described we suggest that it accounts for ∼8–11 mm yr−1 of right-lateral slip. The regular stations distribution across the southern sector of ZFTB allows better detecting the crustal deformation field. Westward of MZRF, the ground deformation field is characterized by a prevailing S-directed motion, passing to an SW-ward motion approaching the frontal trust belt. This rotation is coupled by a progressive reduction of geodetic rates passing from values of ∼10–12 to ∼1–3 mm yr−1, respectively, from the inner side to the external side of ZFTB. By using a simple vectorial decomposition (by grouping GPS velocities on each side of a selected fault), we estimate ∼2.8 and ∼3.4 mm yr−1 of right-lateral motion for the Izeh−Hendijan and Kazerun strike-slip faults, respectively. The prevailing contractional nature of ZFTB is evident on the 2-D geodetic strain-rate map (Fig. 3b). In particular, the maximum contractional horizontal strain-rate shows a fan-shaped feature across the southern sector of ZFTB, maintaining always an orthogonal orientation with respect to the curvature of the collisional mountain belt (‘orocline’); across this area, a shortening up to ∼50 nanostrain per year can be recognized. Moving toward the Turkish–Iranian Plateau, such a prevailing contractional pattern progressively turn into a pattern characterized by an NS shortening up to ∼40 nanostrain per year coupled with an EW extension up to ∼45 nanostrainper year, suggesting a prevailing strike-slip deformation of the Plateau. 3.2 Seismic data Instrumental seismic monitoring of the study area started in 1957 when the Iranian Seismological Centre (http://irsc.ut.ac.ir) deployed a seismograph network. Since then, other regional (e.g. IIEES, http://www.iiees.ac.ir and BHRC, http://www.bhrc.ac.ir) and local seismic networks have been established, thereby increasing the number of seismic stations and improving the quality of earthquake observables (i.e. hypocentral locations and source mechanism determinations). From the International Seismological Centre (ISC, www.isc.ac.uk) online catalogue, we compiled a database of more than 66 000 seismic events occurring since 1909 to date and having focal depth ≤ 50 km and M ≥ 1.0 (Fig. 4a; see also Table S5, Supporting Information section). The ISC routinely produce catalogues of earthquake hypocentre locations relying on data contributed by seismological agencies from around the world. The location provided by the ISC is based on P-wave traveltime tables derived from global 1-D earth velocity model (e.g. the ak135 model; Kennett et al. 1995). The resulting hypocentral parameters are based entirely on reported first-arriving P-wave times, which for most events do not include P-wave arrivals corresponding to upgoing ray paths. Therefore, many ISC hypocentres are poorly constrained in focal depth and must be interpreted with caution. Figure 4. View largeDownload slide (a) Instrumental crustal seismicity (M ≥ 2.5) occurring in the investigated area since 1900 (International Seismological Centre, www.isc.ac.uk). (b) Lower hemisphere, equal area projection for FPSs (with M ≥ 5) compiled from the GCMT catalogue (http://www.globalcmt.org; Dziewonski et al.1981; Ekström et al.2012); FPSs are coloured according to rake: red indicates pure thrust faulting, blue is pure normal faulting and yellow is strike-slip faulting. The inset shows a ternary plot of FPSs: each point is plotted based on the plunge of the P, T and B axes of FPS (Frohlich 1992). The dashed lines divide the diagram into faulting styles based on definitions by Zoback (1992): NF is normal faulting, NS is normal and strike-slip faulting, SS is strike-slip faulting, TS is thrust and strike-slip faulting, TF is thrust faulting and U is undefined faulting. The map is plotted in an oblique Mercator projection. Figure 4. View largeDownload slide (a) Instrumental crustal seismicity (M ≥ 2.5) occurring in the investigated area since 1900 (International Seismological Centre, www.isc.ac.uk). (b) Lower hemisphere, equal area projection for FPSs (with M ≥ 5) compiled from the GCMT catalogue (http://www.globalcmt.org; Dziewonski et al.1981; Ekström et al.2012); FPSs are coloured according to rake: red indicates pure thrust faulting, blue is pure normal faulting and yellow is strike-slip faulting. The inset shows a ternary plot of FPSs: each point is plotted based on the plunge of the P, T and B axes of FPS (Frohlich 1992). The dashed lines divide the diagram into faulting styles based on definitions by Zoback (1992): NF is normal faulting, NS is normal and strike-slip faulting, SS is strike-slip faulting, TS is thrust and strike-slip faulting, TF is thrust faulting and U is undefined faulting. The map is plotted in an oblique Mercator projection. Keeping in mind this main limitation, the instrumental seismicity is mainly distributed on the Turkish–Iranian Plateau and along the Zagros collisional belt (Fig. 4a). In the former, seismicity appears rather spread out, while along the latter, seismicity is confined between the Persian Gulf and the MZRF. The area separating these two regions shows a rather low level of seismicity with sporadic occurrence of large and destructive earthquakes (e.g. the M = 7.2 1962 Buyin Zara earthquake; Fig. 4a), while along the Zagros collisional belt the number of moderate size earthquakes (4.5 ≤ M ≤ 6) is large and no earthquakes with magnitude larger than 7 have been recorded. In order to depict the main seismotectonic features of the investigated area, we compiled a database of focal plane solutions (FPSs) from the Global Centroid Moment Tensor online catalogue (Dziewonski et al.1981; Ekström et al.2012; www.globalcmt.org). FPSs with reverse faulting predominate in Southern Zagros and are characterized by strikes subparallel to the local trend of the topography and fold axes (Fig. 4b). In this area, a number of strike-slip solutions are also evident, mostly associated with the N-S-trending strike-slip right-lateral Kazerun fault and along secondary structures within the Simply Folded Zone. Along Northern Zagros, FPSs are mainly distributed along the external front of the collisional belt and are associated with a few strike-slip solutions occurring along MRF (Fig. 4b). The Turkish–Iranian Plateau is characterized by an equal mixture of reverse and strike-slip solutions. Overall, the FPSs patterns clearly depict the prevailing contractional nature of Southern Zagros (locally segmented along N-S-trending strike-slip faults), which passes to a transpressional faulting regime toward the NW. 4 MOMENT-RATES COMPUTATION AND COMPARISONS In the following, we estimated the scalar geodetic and seismic moment-rates in order to provide an improved evaluation of the seismic/geodetic deformation-rate comparison for the ZFTB. To this aim, we divided the investigated regions into regular 2° × 2° grid (Fig. 3b) in order to have a consistent seismic data set (e.g. catalogue duration and range of magnitude) for each cell. Therefore, in the following we refer to this 2° × 2° grid. The geodetic moment-rate (and its standard error) was estimated for cells of surface area A by adopting the Savage & Simpson (1997) formulation:   $${\dot{M}_{{\rm{geod}}}} = 2\mu {H_s}A[{\rm{Max}}(|{\varepsilon _{H\max }}|,|{\varepsilon _{h\min }}|,|{\varepsilon _{H\max }} + {\varepsilon _{h\min }}|)]$$ (1) where μ is the shear modulus of the rocks (taken here as 3.3·× 1010 N m−2), Hs is the seismogenic thickness over which strains accumulate and its elastic part release in earthquakes, εHmax and εhmin are the principal horizontal strain-rates previously described and Max is a function returning the largest of the arguments. The moment-rate estimate from geodetic strain-rates is proportional to the chosen seismogenic thickness Hs. As mentioned earlier, focal depths for many events in the ISC catalogue cannot be determined with sufficient accuracy. Concerning the ZFTB, uncertainties associated to focal depths are on the order of 10 km, even for the ‘best case’ examples where depth phases (pP, sP and sS) are picked (see Engdahl et al.2006 for additional details). Moreover, a number of recent studies, based on local recordings and teleseismic body-waveform modeling have demonstrated that most of the well-located earthquakes occurs in the 10–14 km depth interval and that most of the events deeper than ∼18 km occur in the south-eastern sector of ZFTB (e.g. Talebian & Jackson 2004; Engdahl et al.2006; Nissen et al.2011; Ansari & Zamari 2014). Hence, based on these considerations, we set Hs= 15 km. For each cell, the seismic moment-rate was calculated according to the Hyndman & Weichert (1983) formulation:   $${\dot{M}_{{\rm{seis}}}} = \phi \frac{b}{{\left( {c - b} \right)}}{10^{\left[ {\left( {c - b} \right){M_x} + a + d} \right]}}$$ (2)which is obtained by integrating the cumulative truncated Gutenberg–Richter distribution up to a maximum magnitude Mx, that is, the magnitude of the largest earthquake that could occur within a specified region. φ is a correction for the stochastic magnitude–moment relation; according to Hyndman & Weichert (1983), we assumed φ = 1.27, reflecting a standard error of 0.2 on magnitudes. c and d are the coefficients of the magnitude (M)–scalar moment (Mseis) relation:   $$\log {M_{{\rm{seis}}}} = cM + d$$ (3) According to Hanks & Kanamori (1979), we set c = 1.5 and d = 9.05. We are aware that the earthquake magnitudes in the ISC catalogue refer to different scales (e.g. mb, body wave magnitude; Ms, surface wave magnitude and MD duration magnitude) which ideally should be converted into moment magnitude (Mw) and use that as a standard, given the limitations of the other magnitude scales. Although Karimiparidari et al. (2013) provide some relationships between Mw and the other magnitude types, we prefer to convert all earthquake magnitudes directly into scalar moments by using the above generalized relation, because in any case, both estimations will always suffer from substantial uncertainties. The coefficients a and b are the seismicity level and the slope of the Gutenberg–Richter recurrence relation, respectively:   $$\log {N_M} = a - bM$$ (4) where, for each cell, NM is the cumulative number of earthquakes of magnitude M and larger. In order to estimate a and b, we adopted the maximum likelihood method (Weichert 1980). Both coefficients have been primarily constrained by the small- to mid-size earthquakes (M < 5), using as input the instrumental catalogue. In some few cells (Z008 and Z017), externally located with respect to ZFTB, the number of earthquakes is relatively small (<50) and can be regarded as a poor constraint, especially on the b value estimation. Hence, results obtained for these cells have been omitted. Results achieved for cells containing lesser than 200 earthquakes (Z006, Z016, Z021, Z022, Z024 and Z030) can be viewed as adequate, while those achieved for cells containing more than 200 earthquakes can be considered from good to highly reliable. Estimated parameters are reported in Table S6 of the Supporting Information section. According to eq. (2), the seismic moment-rate estimation for each cell depends by the Mx value. A simple method of calculating Mx is to use the largest earthquake in the historical catalogue and add 0.5 (e.g. Kijko & Graham 1998), however such a method is very limited where there is no significant historical record. Another method is to use scaling relations between the length of the fault and the maximum earthquake (e.g. Wells & Coppersmith 1994). This method can be applied where there are no historical data, but a number of issues come with deciding on whether, and how, to divide the fault up into segments. As alternative, Mx can be estimated by using statistical approaches (see also the Supporting Information section). In this study, we estimated the Mx value by using the MMAX toolbox developed by Kijko & Singh (2011). Such a toolbox, by adopting a wide spectrum of statistical procedures, allows to estimate the Mx values for a given area, in different circumstances (completeness and temporal length of the catalogue, magnitude distribution and uncertainties, number of earthquakes, etc; see Kijko & Singh 2011 for details). For each cell, we estimated the Mx values according with the 12 procedures implemented in the toolbox and using as input two different data sets. In a former step, we used as input the instrumental catalogue which spans the 1909–2016 time interval; the main results are reported in Table S7 in the Supporting Information. In this step, a standard error of 0.25 in magnitude determination of all events have been considered. In a latter step, we used as input a catalogue including both the instrumental catalogue used in the former computation and the historical seismic records (Mirzaei et al.1997; Ambraseys & Jackson 1998); the main results are reported in Table S8 in the Supporting Information. Since magnitude uncertainties in the historical catalogue are hard to measure and quantify (and usually increase further back in time) in this step, a standard error of 0.35 in magnitude determination of all events have been considered. Considering all the results coming from the two steps, the estimations based only on the instrumental data usually underestimate the Mx values, since the ‘seismic cycle’ over the investigated area is longer than the duration of the instrumental record time span. On the other hand, the estimations based on the mixed catalogue (covering the last 1200 yr) agree with the highest observed magnitudes, reflecting however, a strictly dependence of Mx estimations to the historical records (e.g. the number of historic events over higher magnitude ranges in each cell, event magnitude uncertainties). Both these features can be clearly observed from the cumulative frequency–magnitude plots reported in Fig. S6 in the Supporting Information: for instance, the estimated 7.84 Mx value (Tables S8, Supporting Information; the estimation is based on two M = 7.7 earthquakes, whose estimated magnitude are affected by large uncertainties) for cell Z006, is hardly derivable from the plot, while a value of 5.8 (Table S7, Supporting Information; no earthquakes with M > 6 occurred in the cell during the instrumental period) should be more appropriate. Furthermore, it must be also noted that, the Mx value has a significant effect on the seismic moment-rates estimation (2); an increase of 1 mag unit leads to an increase of the moment-rate by a factor of ∼5.4 (e.g. Mazzotti et al.2011). Keeping in mind all these aspects, in the following, we refer to the estimations based on the mixed catalogue. The toolbox does not indicate which of the resulting estimations is the more appropriate; hence, for each cell we selected the one having the smallest uncertainties (see the Supporting Information section and Fig. S7 for details). Finally, for each cell, the seismic moment-rates (and associated uncertainties) have been estimated according to the relationship (2), by taking into account all the parameters as previously defined/computed. Uncertainties in our estimates have been provided as a median value along with an associated 67 per cent confidence interval, by adopting a logic tree approach, similar to the one described in Mazzotti & Adams (2005). The seismic moment-rate can be estimated also by adopting the Kostrov (1974) formulation:   $${\dot{M}_{{\rm{seis}}}} = \frac{1}{{\Delta T}}\sum\limits_{n = 1}^N {M_{{\rm{seis}}}^{(n)}}$$ (5) where N is the number of events occurring in the volume A · Hs (A and Hs represent the surface area and the seismogenic thickness previously defined, respectively), Mseis is the seismic moment of the nth earthquake from the N total earthquakes and ΔT represents the time interval considered. We mainly focused on the results obtained from the truncated Gutenberg–Richter distribution, however, in order to test sensitivity, we additional computed the Kostrov summation (see the Supporting Information section for details). The estimated geodetic and seismic moment-rates are both reported in Fig. 5 and Table 1. Regarding the geodetic moment-rates, we performed an additional computation by taking into account the strain-rate field estimated with the method described in Shen et al. (2015) and reported in Fig. S1a in the Supporting Information. Both geodetic moment-rate estimates (reported in Fig. S8, Supporting Information), with the exception of very few cells (e.g. Z002, Z015, Z020 and Z029), well overlaps within their associated uncertainties. This fact clearly highlights as the crustal pattern of deformation over the investigated area is adequately sampled by both methods, although they rely on different approaches in velocity gradient computation. The following considerations and computations are primarily based on the results coming from the ‘spline in tension’ method, with references to the alternative results where they are significant. Figure 5. View largeDownload slide Estimated moment-rates from (a) geodetic data, (b) instrumental seismicity, by considering the truncated Gutenberg–Richter distribution and (c) instrumental seismicity, by considering the cumulative Kostrov summation method (see Table 1 for details). The maps are plotted in an oblique Mercator projection. Figure 5. View largeDownload slide Estimated moment-rates from (a) geodetic data, (b) instrumental seismicity, by considering the truncated Gutenberg–Richter distribution and (c) instrumental seismicity, by considering the cumulative Kostrov summation method (see Table 1 for details). The maps are plotted in an oblique Mercator projection. Table 1. Moment-rates (N·m yr−1) estimated from geodetic data (GM1) and instrumental seismicity estimated by integrating the cumulative truncated Gutenberg–Richter distribution up to a maximum magnitude Mx (SM–GR) and by adopting the cumulative Kostrov summation method (SM–K). Cell ID  Long.  Lat.  GM1  SM–GR  SM–K  SCC–GR  SCC–K        (1017 N·m yr−1)  (1017 N·m yr−1)  (1017 N·m yr−1)  (per cent)  (per cent)  Z001  43.000  39.000  16.2 ± 1.2  $$13.1_{ - 4.1}^{ + 7.1}$$  3.1  $$80.7_{ - 20.7}^{ + 35.1}$$  18.9  Z002  45.000  39.000  17.5 ± 1.4  $$5.3_{ - 2.4}^{ + 6.6}$$  9.0  $$30.1_{ - 11.1}^{ + 30.0}$$  51.6  Z005  45.000  37.000  12.1 ± 2.2  $$2.7_{ - 1.2}^{ + 3.4}$$  0.1  $$22.2_{ - 8.8}^{ + 22.3}$$  0.8  Z006  47.000  37.000  7.4 ± 0.6  $$6.6_{ - 2.6}^{ + 5.7}$$  0.01  $$88.6_{ - 28.8}^{ + 62.3}$$  0.6  Z007  49.000  37.000  6.3 ± 2.1  $$8.7_{ - 4.3}^{ + 9.9}$$  1.1  $$138.3_{ - 66.1}^{ + 132.0}$$  16.8  Z011  49.000  35.000  7.2 ± 1.7  $$6.6_{ - 2.8}^{ + 8.2}$$  12.9  $$91.2_{ - 39.3}^{ + 106.1}$$  179.4  Z012  51.000  35.000  8.64 ± 2.1  $$2.4_{ - 1.3}^{ + 4.5}$$  0.1  $$27.4_{ - 12.7}^{ + 42.2}$$  1.6  Z013  45.000  33.000  5.0 ± 1.6  $$1.6_{ - 0.7}^{ + 1.8}$$  0.1  $$31.4_{ - 13.8}^{ + 30.1}$$  2.1  Z014  47.000  33.000  7.8 ± 3.1  $$6.3_{ - 2.5}^{ + 5.4}$$  1.0  $$81.2_{ - 35.9}^{ + 60.9}$$  12.2  Z015  49.000  33.000  12.1 ± 2.0  $$2.5_{ - 1.0}^{ + 2.2}$$  0.8  $$20.5_{ - 7.2}^{ + 15.0}$$  6.8  Z016  51.000  33.000  6.9 ± 2.2  $$0.7_{ - 0.4}^{ + 2.2}$$  0.02  $$9.6_{ - 5.2}^{ + 26.0}$$  0.3  Z019  49.000  31.000  9.4 ± 2.6  $$5.5_{ - 2.1}^{ + 4.5}$$  0.2  $$58.2_{ - 22.0}^{ + 39.9}$$  2.6  Z020  51.000  31.000  12.5 ± 2.1  $$4.5_{ - 2.3}^{ + 7.6}$$  1.0  $$35.9_{ - 15.2}^{ + 48.7}$$  7.6  Z021  53.000  31.000  4.8 ± 2.6  $$1.8_{ - 0.9}^{ + 3.4}$$  0.1  $$37.8_{ - 23.1}^{ + 66.0}$$  1.1  Z022  55.000  31.000  3.4 ± 1.9  $$0.1_{ - 0.0}^{ + 0.1}$$  0.3  $$1.7_{ - 1.0}^{ + 3.2}$$  8.4  Z023  57.000  31.000  6.9 ± 1.6  $$3.2_{ - 1.2}^{ + 2.5}$$  0.7  $$46.9_{ - 16.8}^{ + 30.5}$$  10.8  Z025  51.000  29.000  13.7 ± 3.0  $$2.9_{ - 1.0}^{ + 1.9}$$  2.2  $$21.4_{ - 6.9}^{ + 11.8}$$  15.7  Z026  53.000  29.000  17.1 ± 1.3  $$5.7_{ - 3.3}^{ + 16.4}$$  1.1  $$33.3_{ - 15.4}^{ + 76.3}$$  6.1  Z027  55.000  29.000  12.9 ± 2.0  $$0.9_{ - 0.4}^{ + 1.4}$$  0.7  $$7.1_{ - 2.9}^{ + 8.6}$$  5.5  Z028  57.000  29.000  13.8 ± 2.1  $$4.6_{2.1}^{ + 6.0}$$  2.6  $$33.5_{ - 13.0}^{ + 35.1}$$  18.5  Z029  59.000  29.000  6.2 ± 0.6  $$1.9_{ - 1.1}^{ + 5.1}$$  0.7  $$31.3_{ - 14.1}^{ + 65.6}$$  11.4  Z031  53.000  27.000  15.4 ± 2.7  $$2.3_{ - 1.0}^{ + 2.3}$$  0.6  $$15.1_{ - 5.3}^{ + 11.9}$$  3.7  Z032  55.000  27.000  16.9 ± 0.7  $$4.7_{ - 2.0}^{ + 5.3}$$  2.4  $$27.5_{ - 9.6}^{ + 25.1}$$  14.5  Z033  57.000  27.000  21.6 ± 3.4  $$4.4_{ - 1.6}^{ + 3.1}$$  1.7  $$20.5_{ - 6.3}^{ + 11.9}$$  7.8  Cell ID  Long.  Lat.  GM1  SM–GR  SM–K  SCC–GR  SCC–K        (1017 N·m yr−1)  (1017 N·m yr−1)  (1017 N·m yr−1)  (per cent)  (per cent)  Z001  43.000  39.000  16.2 ± 1.2  $$13.1_{ - 4.1}^{ + 7.1}$$  3.1  $$80.7_{ - 20.7}^{ + 35.1}$$  18.9  Z002  45.000  39.000  17.5 ± 1.4  $$5.3_{ - 2.4}^{ + 6.6}$$  9.0  $$30.1_{ - 11.1}^{ + 30.0}$$  51.6  Z005  45.000  37.000  12.1 ± 2.2  $$2.7_{ - 1.2}^{ + 3.4}$$  0.1  $$22.2_{ - 8.8}^{ + 22.3}$$  0.8  Z006  47.000  37.000  7.4 ± 0.6  $$6.6_{ - 2.6}^{ + 5.7}$$  0.01  $$88.6_{ - 28.8}^{ + 62.3}$$  0.6  Z007  49.000  37.000  6.3 ± 2.1  $$8.7_{ - 4.3}^{ + 9.9}$$  1.1  $$138.3_{ - 66.1}^{ + 132.0}$$  16.8  Z011  49.000  35.000  7.2 ± 1.7  $$6.6_{ - 2.8}^{ + 8.2}$$  12.9  $$91.2_{ - 39.3}^{ + 106.1}$$  179.4  Z012  51.000  35.000  8.64 ± 2.1  $$2.4_{ - 1.3}^{ + 4.5}$$  0.1  $$27.4_{ - 12.7}^{ + 42.2}$$  1.6  Z013  45.000  33.000  5.0 ± 1.6  $$1.6_{ - 0.7}^{ + 1.8}$$  0.1  $$31.4_{ - 13.8}^{ + 30.1}$$  2.1  Z014  47.000  33.000  7.8 ± 3.1  $$6.3_{ - 2.5}^{ + 5.4}$$  1.0  $$81.2_{ - 35.9}^{ + 60.9}$$  12.2  Z015  49.000  33.000  12.1 ± 2.0  $$2.5_{ - 1.0}^{ + 2.2}$$  0.8  $$20.5_{ - 7.2}^{ + 15.0}$$  6.8  Z016  51.000  33.000  6.9 ± 2.2  $$0.7_{ - 0.4}^{ + 2.2}$$  0.02  $$9.6_{ - 5.2}^{ + 26.0}$$  0.3  Z019  49.000  31.000  9.4 ± 2.6  $$5.5_{ - 2.1}^{ + 4.5}$$  0.2  $$58.2_{ - 22.0}^{ + 39.9}$$  2.6  Z020  51.000  31.000  12.5 ± 2.1  $$4.5_{ - 2.3}^{ + 7.6}$$  1.0  $$35.9_{ - 15.2}^{ + 48.7}$$  7.6  Z021  53.000  31.000  4.8 ± 2.6  $$1.8_{ - 0.9}^{ + 3.4}$$  0.1  $$37.8_{ - 23.1}^{ + 66.0}$$  1.1  Z022  55.000  31.000  3.4 ± 1.9  $$0.1_{ - 0.0}^{ + 0.1}$$  0.3  $$1.7_{ - 1.0}^{ + 3.2}$$  8.4  Z023  57.000  31.000  6.9 ± 1.6  $$3.2_{ - 1.2}^{ + 2.5}$$  0.7  $$46.9_{ - 16.8}^{ + 30.5}$$  10.8  Z025  51.000  29.000  13.7 ± 3.0  $$2.9_{ - 1.0}^{ + 1.9}$$  2.2  $$21.4_{ - 6.9}^{ + 11.8}$$  15.7  Z026  53.000  29.000  17.1 ± 1.3  $$5.7_{ - 3.3}^{ + 16.4}$$  1.1  $$33.3_{ - 15.4}^{ + 76.3}$$  6.1  Z027  55.000  29.000  12.9 ± 2.0  $$0.9_{ - 0.4}^{ + 1.4}$$  0.7  $$7.1_{ - 2.9}^{ + 8.6}$$  5.5  Z028  57.000  29.000  13.8 ± 2.1  $$4.6_{2.1}^{ + 6.0}$$  2.6  $$33.5_{ - 13.0}^{ + 35.1}$$  18.5  Z029  59.000  29.000  6.2 ± 0.6  $$1.9_{ - 1.1}^{ + 5.1}$$  0.7  $$31.3_{ - 14.1}^{ + 65.6}$$  11.4  Z031  53.000  27.000  15.4 ± 2.7  $$2.3_{ - 1.0}^{ + 2.3}$$  0.6  $$15.1_{ - 5.3}^{ + 11.9}$$  3.7  Z032  55.000  27.000  16.9 ± 0.7  $$4.7_{ - 2.0}^{ + 5.3}$$  2.4  $$27.5_{ - 9.6}^{ + 25.1}$$  14.5  Z033  57.000  27.000  21.6 ± 3.4  $$4.4_{ - 1.6}^{ + 3.1}$$  1.7  $$20.5_{ - 6.3}^{ + 11.9}$$  7.8  Notes: Values of 15 km and 33 GPa have been used, respectively, for the seismogenic thickness and rigidity modulus for GM1 computation. See the main text on adopted formulation. Seismic coupling coefficient (SCC) expressed as percentage of the seismic/geodetic moment-rate ratio is also reported. Uncertainties (67 per cent confidence interval) of estimated moment-rates and SCC are also reported; confidence intervals are strongly asymmetric, with large upper bounds, due to the asymmetry of the moment-rate lognormal distribution (2). View Large Estimated geodetic moment-rates (Table 1 and Fig. 5a) range in the interval 3.4 × 1017-2.2 × 1018 N·m yr−1. The higher values (≥1.5 × 1018 N·m yr−1) are observed in Southern Zagros (along the Fars arc and close to the Oman line; e.g. Z026, Z031, Z232 and Z033) and close to the Urmia and Van lakes region (Turkish–Iranian Plateau; e.g. Z001 and Z002). Seismic moment-rates (Table 1 and Fig. 5b), estimated by adopting the truncated Gutenberg–Richter relation, range in the interval 5.6 × 1015-1.3 × 1018 N·m yr−1. The higher values of the seismic moment-rates (≥1 × 1018 N·m yr−1) are observed only along the Turkish–Iranian Plateau (Z001). The Turkish–Iranian Plateau as well as Northern Zagros and the Sanandaj–Sirjan zone are also characterized by some cells with values ranging in the 5 × 1017-1 × 1018 N·m yr−1 interval (Z002, Z006, Z007, Z011, Z014 and Z019; Fig. 5b). Southern Zagros are prevailing characterized by cells with values ranging in the 2 × 1017-5 × 1017 N·m yr−1 interval, while the lower values (<1.× 1017 N·m yr−1) are observed on some cells located along the Sanandaj–Sirjan zone (Z016, Z022 and Z027; Fig. 5b). Seismic moment-rates, computed by adopting the Kostrov summation method, range in the interval 1.9 × 1015-1.3 × 1018 N·m yr−1 (Table 1 and Fig. 5b). The higher values are observed in only two cells, located nearby the Urmia lake (Z002 with a cumulative moment-rate of 9.0 × 1017 N·m yr−1) and the Buyin Zara (Z011 with a cumulative moment-rate of 1.3 × 1018 N·m yr−1) regions, respectively, which have been both characterized by the occurrence of destructive earthquakes (M > 7) during the last century (see Fig. 4a). Some cells located nearby the Van lake (Z001), the Buyin Zara zone (Z007) and on Southern Zagros (Z025, Z026, Z028, Z032 and Z033; Fig. 5b) are characterized by moment-rates ranging in the 1.0 × 1017–3.1.× 1017 N·m yr−1 interval. In the remaining sectors of the investigated area, the seismic moment-rates show values lower than 1.0 × 1017 N·m yr−1. Considering that the geodetic moment-rate is a measure of both elastic and anelastic loading rates, while the seismic moment-rate is a measure of the elastic unloading rate, the simple seismic/geodetic moment-rate ratio, expressed as a percentage, can be termed ‘seismic coupling coefficient’ (hereinafter SCC). The higher this ratio, the larger part of the measured deformation has been seismically released. Conversely, a low ratio indicates an apparent seismic moment deficit, which suggests either a proportion of aseismic deformation (i.e. ongoing unloading by creep and other plastic process) or overdue earthquakes (i.e. elastic storage). Fig. 6(a) shows the distribution of the SCC over the investigated area (estimated values are reported in Table 1; see also Fig. S9, Supporting Information). On the Turkish–Iranian Plateau and along the Sanandaj–Sirjan zone, the estimated SCC values range in the intervals 22–89 and 1–91 per cent, respectively. On Northern and Southern Zagros, the SCC values range in the intervals 20–81 and 15–33 per cent, respectively. On Fig. 6(b), we reported the SCC values resulting from the seismic/geodetic moment-rate ratio with seismic moment-rates estimated by adopting the Kostrov approach. The SCC value exceeds 50 per cent only on the two cells (Z002 and Z011), which, as previously described, have been affected by the occurrence of destructive earthquake (M > 7) during the last century. On the other cells, the SCC pattern does not exceed the value of 19 per cent (∼8 per cent on average). Figure 6. View largeDownload slide Seismic coupling coefficient (SCC), expressed as percentage of the seismic/geodetic moment-rate ratio as computed in this study (see Table 1 for details). (a) SCC from seismic moment-rates derived by the truncated Gutenberg–Richter distribution. The dotted red line marks the boundary of the evaporitic Hormuz Formation (e.g. Bahroudi & Koyi 2003). (b) SCC from seismic moment-rates derived by the cumulative Kostrov summation method. The maps are plotted in an oblique Mercator projection. Figure 6. View largeDownload slide Seismic coupling coefficient (SCC), expressed as percentage of the seismic/geodetic moment-rate ratio as computed in this study (see Table 1 for details). (a) SCC from seismic moment-rates derived by the truncated Gutenberg–Richter distribution. The dotted red line marks the boundary of the evaporitic Hormuz Formation (e.g. Bahroudi & Koyi 2003). (b) SCC from seismic moment-rates derived by the cumulative Kostrov summation method. The maps are plotted in an oblique Mercator projection. 5 DISCUSSION AND CONCLUSIONS An extensive combination of novel observations rigorously integrated with already published results improves the picture of the ongoing crustal deformation field for the entire ZFTB. Taking into account the instrumental seismicity catalogues, we have been able to provide a statistical evaluation of the seismic/geodetic deformation-rate ratio (i.e. SCC) for the area. Our GPS-based velocity field, referred to the Arabian plate, allows recognizing a clockwise rotation, passing from a ∼11–13 mm yr−1 SE-ward motion along the Turkish–Iranian Plateau to a ∼10–12 mm yr−1 SW-directed motion on the inner side of Fars arc. Sites located along the external margin of ZFTB show ∼1–3 mm yr−1 of motion, highlighting how much of the oblique Arabia–Eurasia convergence is currently absorbed within the Zagros with prevailing contractional features across Southern Zagros and right-lateral shear along the NW trending MRF in NW Zagros. Due to convergence rate variations along the general NW–SE strike of Zagros, the oblique collision is transformed by some roughly N-S-trending strike-slip faults. The dense coverage of our GPS solution allowed us to estimate ∼2.8 mm yr−1 of right-lateral motion for the Izeh–Hendijan fault and ∼3.4 mm yr−1 of right-lateral motion for the Kazerun fault (Figs 2a and 3a). No geological slip-rate estimations are currently available for the Izeh–Hendijan fault, while our estimation for the Kazerun fault matches well with previous geodetic (∼3.6 mm yr−1; Tavakoli et al.2008) and geological estimations (1.5–4 mm yr−1; Authemayou et al.2009, and references therein). These faults have a strike parallel to the Eurasia–Arabia convergence direction and, while partitioning a fraction of the convergence rate, segment the Southern Zagros range along the prevailing NS direction, as already evidenced by previous studies (e.g. Authemayou et al.2009, and references therein). We also suggest that MRF accounts for ∼8–11 mm yr−1 of right-lateral slip-rate. The right-lateral strike-slip faulting along the MRF is confirmed by focal mechanism solutions of earthquakes and tectonic geomorphology (see Talebian & Jackson 2004, and references therein). Alipoor et al. (2012) evaluated a slip-rate of 1.6–3.2 mm yr−1 along the MRF by using geological, geomorphological markers and drainage patterns. Based on GPS measurements, Vernant et al. (2004) and Walpersdorf et al. (2006) estimated 3 ± 2 and 4–6 mm yr−1 of right-lateral slip-rate for the MRF, respectively. Copley & Jackson (2006) and Authemayou et al. (2009) also suggested a slip-rate of 2–5 and 3.5–12.5 mm yr−1 along the MRF and close to its northwestern termination, respectively. As above mentioned, the lack of GPS measurements close to the MRF does not allow direct measurement of the geodetic slip-rate on this fault system. However, although our estimated rate is based on a simple vectorial decomposition (by grouping GPS velocities on each side of the fault), it concurs with the upper limit of the rates determined from long-term geomorphic offsets by Authemayou et al. (2006). The map of the SCC over the investigated area, while confirms some first-order features established by previous estimations (e.g. Jackson & McKenzie 1988; Masson et al.2005; Ansari & Zamani 2014; Khodaverdian et al.2015), leads also to some further considerations. It must be noted that previous estimates have adopted different (i) data sets, (ii) approaches and (iii) subdivision of the investigated area. The above-quoted authors, used both instrumental and historical seismicity catalogues (with different temporal coverage and magnitude threshold) and usually adopted the Kostrov approach by subdividing the area into large polygons. In addition, also the used geodetic data sets show different spatial coverages, passing from a sparse geodetic network as in Masson et al. (2005) to a dense network as in Ansari & Zamani (2014) and Khodaverdian et al. (2015). Our moment-rates and CSS estimations are based on the densest geodetic data set currently available for the investigated area and a seismic catalogue (from the ISC) spanning the 1909–2016 time interval. We are aware (i) that the ISC catalogue contains a substantial number of earthquakes which have been determined from teleseismic waveforms and that, for many of them, the location is insufficiently accurate (with uncertainties of up to 50 km; Talebian & Jackson 2004) and (ii) that such a catalogue can be considered complete for M ≥ 4.4 since 1960 (Karimiparidari et al.2013). While the uncertainty related to the poorly accurate earthquake locations cannot be properly addressed, the use of the cumulative truncated Gutenberg–Richter earthquake distribution, allow us to account for the probable incompleteness of the seismic catalogue. On the Turkish–Iranian Plateau, the SCC pattern is sampled only by four cells (Fig. 6a). Cells Z001 and Z006 are characterized by SCC values >80 per cent, while cells Z002 and Z005 show values of 30 and 22 per cent (or 60 and 33 per cent by taking into account the Shen et al. (2015) strain-rate field), respectively. On the basis of the earthquakes statistics, results achieved for cells Z001, Z002 and Z005 can be considered from good to highly reliable since them are well constrained by geodetic and seismological observations, while results achieved for cell Z006 can be viewed as adequate. Overall, achieved results well testify, beside the low SCC values inferred for cell Z005 (which have experienced only three M > 6 earthquakes in the last 1200 yr), that on the Turkish–Iranian Plateau, a large fraction (>60 per cent) of the measured crustal deformation occurs seismically. On the Sanandaj–Sirjan zone, the SCC pattern is mainly sampled by seven cells (Fig. 6a). The highest value has been inferred for cell Z011 (∼91 per cent), while, with the exception of some solitary cells (Z016, Z022 and Z027) characterized by values <10 per cent, the remaining cells (Z012, Z021 and Z028) show values in the 27–38 per cent range. Results achieved for cell Z011, although well constrained by earthquakes statistics, would be lesser reliable because the low number of GPS stations and/or the short temporal observation period (4–6 yr), therefore providing an inaccurate sample of the ongoing deformation pattern. Therefore, by excluding results achieved for cell Z011, we suggest that on Sanandaj–Sirjan zone a small to moderate fraction (<40 per cent) of the measured crustal deformation occurs seismically. On some cells (e.g. Z016 and Z027), the occurrence of large earthquakes in the past (estimated magnitudes M ≥ 6) coupled with a high geodetic deformation, could indicate overdue M ≥ 6 earthquakes. Similar considerations can be done by taking into account the moment-rates computed by using the Shen et al. (2015) strain-rate estimates (Figs S8 and S9, Supporting Information). Cells located eastward the Sanandaj–Sirjan zone, are characterized by moderate SCC values (31 and 47 per cent for cells Z023 and Z029, respectively), or by SCC values >100 per cent (cell Z007). The formers show a seismic behaviour similar to the one inferred for the Sanandaj–Sirjan zone, the latter, beside the apparent imbalance between the proxies of moment-rate, clearly highlight how a high fraction of the measured crustal deformation occurs seismically. On Northern Zagros, the SCC pattern is sampled by five cells (Fig. 6a). Cells Z014 and Z019 are characterized by SCC values of 81 and 52 per cent, respectively, while the remaining cells (Z013, Z015 and Z020) show values in the 20–36 per cent range. Results achieved for cell Z019 would be lesser reliable because Mx is constrained by only a large earthquake (M = 6.5, 840 AD; earthquakes occurred during the instrumental time do not exceed M = 5.5) whose magnitude estimation is poorly constrained (Ambraseys & Jackson 1998). This aspect suggests a possible overestimation of SCC value for cell Z019. Results achieved for cell Z014 can be considered lesser reliable also because the low number of GPS stations (that are also affected by large uncertainties) which would provide an inaccurate sample of deformation field over the investigated cell. Conversely, results achieved for cells Z015 and Z020 can be considered from good to highly reliable, since them are well constrained by earthquakes statistics and geodetic data. Considering the moment-rate values computed by using the Shen et al. (2015) strain-rate estimates (Figs S8 and S9, Supporting Information), some differences can be observed. More in detail, these last estimates are generally lower than those estimated by taking into account the ‘spline in tension’ method, leading to an increase of estimated SCC values. This aspect suggests that on this area, some local fluctuations of the strain-rate field would be better captured by the Shen et al. (2015) approach, therefore providing a finer resolution of ongoing deformation. Based on these considerations, we suggest that on Northern Zagros a moderate fraction (∼49 per cent) of the measured crustal deformation occurs seismically. On Southern Zagros, the SCC pattern is mainly sampled by five cells (Fig. 6a). Such a pattern is quite homogeneous with values ranging in the 15–33 per cent interval. On the basis of the earthquakes statistics, values estimated for cell Z026 could be considered less reliable, while for the remaining cells (Z025, Z031, Z032 and Z033) achieved results can be considered as highly reliable. Similar consideration can be done by taking into account moment-rates computed by using the Shen et al. (2015) strain-rate estimates (Figs S8 and S9, Supporting Information). These results well highlight as on Southern Zagros a large fraction of the measured crustal deformation occurs aseismically. Generally, the estimated SCC values concur well with the ones proposed by Khodaverdian et al. (2015), whereas are higher than the ones reported in the previous estimations (Jackson & McKenzie 1988; Masson et al.2005; Ansari & Zamani 2014), especially for the Zagros area. In particular, for this last area, previous estimations do not exceed the value of 15 per cent (see Ansari & Zamani 2014, and references therein). Masson et al. (2005), in estimating the SCC for the whole Iranian region, used different time intervals and observed relatively constant values for the last 100, 200 and 300 yr, concluding that an interval of 200–300 yr better depicts the SCC pattern of this area. By considering our estimations using the Kostrov approach (see the Supporting Information section for additional details), we note a general reduction of the SCC values over the investigated area. This aspect suggests that seismic moment-rates estimated with the Kostrov approach could be underestimated. Indeed, the use of the Kostrov approach suffers from the possible lack of both large earthquakes (with long recurrence interval compared to the catalogue duration) and undetected small magnitude events. As previously mentioned, the use of a statistical method such as the cumulative truncated Gutenberg–Richter earthquake distribution allows to take into account the probable incompleteness of the existing catalogue. Hence, seismic moment-rates calculated with this last approach can be assumed as representative of the long-term seismic deformation over a given region (e.g. Mazzotti et al.2011). Northern and Southern Zagros are characterized by SCC values which, based on the significant differences on the strain-rate patterns above discussed, partially overlaps. However, by taking into account the pattern of estimated seismic and geodetic moment-rates, some interesting features can be observed (Fig. 6). More in detail, while estimated seismic moment-rates are quite similar for both regions, with slight larger values inferred for Northern Zagros, the pattern of geodetic moment-rates strongly differs, with the larger values (≥1.3 × 1018 N·m yr−1) detected on the Southern Zagros (Fig. S8, Supporting Information). These features highlight that in spite of a quite similar seismic release of both regions, Southern Zagros account for the larger crustal deformation, clearly evidencing the prevailing aseismic behaviour of that region. To explain such an aseismic behaviour, a number of different hypotheses have been proposed in the last decade. For instance, Barnhart & Lohman (2013) and Barnhart et al. (2013) by using InSAR data suggested that aseismic shortening is segmented by fault creep in the sedimentary cover. Nissen et al. (2011) argued that the aseismic shortening is concentrated within the basement, as implied by shallow earthquake centroid depths. Similar conclusions have been obtained also by Allen et al. (2013) by comparing earthquake locations and topography. Others suggest that faulting in the basement occurs independently of the folding of the sedimentary cover because the cover is detached from the basement along the weak evaporitic Hormuz Formation resulting in a splitting of the seismogenic layer in two. This might explain the large number of M = 5–6 earthquakes and the complete absence of any greater than M = 7 from both instrumental and historical records (Nissen et al.2014). These observations, suggest that the seismicity pattern observed on Southern Zagros would be primarily driven by buoyancy forces arising from large lateral variations in the density and structure of the lower crust/lithosphere. A large fraction of this seismicity occurs on the upper basement and lower/mid-sedimentary cover, as suggested by the largest reverse faulting earthquakes occurred in the area (e.g. 1972 April 10 and 1977 March 21; see Nissen et al.2014, and references therein). In fact, these large earthquakes, which are located at a depth of ∼10 km and are coupled with the lack of surface rupturing, probably extended across the Hormuz Formation between the basement and the sedimentary cover, because of their obvious larger source dimensions. In addition, the incompetence of the weak evaporitic Hormuz Formation allows the occurrence of large aseismic motion on subhorizontal faults and surfaces of décollement (as suggested also by other authors; e.g. Talbot & Alavi 1996; Mouthereau et al.2007; Jahani et al.2009) hence resulting in a low SCC. Unfortunately, the possibility of splitting of the seismogenic layer in two as proposed by some investigators (Nissen et al.2014) cannot be ruled out by our data. Conversely, the deformation occurring on the Turkish–Iranian Plateau is primary driven by plate stresses which lead to a clear seismic behaviour of the area, resulting into high SCC values. In conclusion, our interpretation of geodetic and seismic data presented in this study might contribute to improve the picture of the current seismotectonic pattern of the Zagros area, clearly indicating that this active area accounts for the large deformation-rates related to the oblique Arabia–Eurasia convergence. In addition, despite a number of uncertainties, mainly related to available historical and instrumental seismic information, we constrain the aseismic fraction of the total deformation-rate budget. Estimated values clearly demonstrate how the largest fraction of aseismic deformation occurs in Southern Zagros. This is corroborating the role of massive salt deposits (the evaporitic Hormuz Formation) as possible décollement in the Southern Zagros crust (e.g. Hatzfeld & Molnar 2010, Nissen et al.2014). Aseismic deformation has generally been recognized in tectonic areas characterized by low-magnitude background seismicity (see Mazzotti et al.2011 for an overview). However, considering the high seismic activity and the large deformation-rates observed in the investigated area, an improved knowledge about the rheological behaviour of the lithosphere is required. Acknowledgements We acknowledge EUREF (www.epncb.oma.be), SOPAC (http://sopac.ucsd.edu/), UNAVCO (www.unavco.org) and NGS (http://geodesy.noaa.gov) for providing free access to GPS data. We wish to thank the Editor, Prof. Duncan Agnew, and two anonymous reviewers for their very constructive comments and suggestions, which helped us to significantly improve the early version of paper. REFERENCES Alipoor R., Zaré M., Ghassemi M.R., 2012. Inception of activity and slip rate on the Main Recent Fault of Zagros Mountains, Iran, Geomorphology , 175–176, 86– 97. https://doi.org/10.1016/j.geomorph.2012.06.025 Google Scholar CrossRef Search ADS   Allen M.B., Saville C., Blanc E.J.-P., Talebian M., Nissen E., 2013. Orogenic plateau growth: expansion of the Turkish-Iranian Plateau across the Zagros fold-and-thrust belt, Tectonics , 32( 2), 171– 190. https://doi.org/10.1002/tect.20025 Google Scholar CrossRef Search ADS   Ambraseys N.N., 1989. Temporary seismic quiescence: SE Turkey, Geophys. J. Int. , 96( 2), 311– 331. https://doi.org/10.1111/j.1365-246X.1989.tb04453.x Google Scholar CrossRef Search ADS   Ambraseys N.N., Jackson J.A., 1998. Faulting associated with historical and recent earthquakes in the Eastern Mediterranean region, Geophys. J. Int. , 133( 2), 390– 406. https://doi.org/10.1046/j.1365-246X.1998.00508.x Google Scholar CrossRef Search ADS   Ambraseys N.N., Melville C., 1982. A History of Persian Earthquakes , Cambridge University Press, Cambridge. Ansari S., Zamari A., 2014. Short-term seismic crustal deformation of Iran, Ann. Geophys. , 57( 2), S0210, doi:10.4401/ag-6413. Aqrawi A.A.M., Goff J.C., Horbury A.D., Sadooni F.N., 2010. The Petroleum Geology of Iraq , Scientific Press Ltd. ArRajehi A. et al.  , 2010. Geodetic constraints on present-day motion of the Arabian Plate: implications for Red Sea and Gulf of Aden rifting, Tectonics , 29 TC3011, 10.1029/2009TC002482. https://doi.org/doi:10.1029/2009TC002482 Google Scholar CrossRef Search ADS   Authemayou C., Chardon D., Bellier O., Malekzadeh Z., Shabanian F., Abbassi M.R., 2006. Late Cenozoic partitioning of oblique plate convergence in the Zagros fold-and-thrust belt (Iran), Tectonics , 25( 3), TC3002, doi:10.1029/2005tc001860. https://doi.org/10.1029/2005TC001860 Google Scholar CrossRef Search ADS   Authemayou C. et al.  , 2009. Quaternary slip-rates of the Kazerun and the Main Recent Faults: active strike-slip partitioning in the Zagros fold-and-thrust belt, Geophys. J. Int. , 178( 1), 524– 540. https://doi.org/10.1111/j.1365-246X.2009.04191.x Google Scholar CrossRef Search ADS   Bahroudi A., Koyi H.A., 2003. Effect of spatial distribution of Hormuz salt on deformation style in the Zagros fold and thrust belt: an analogue modelling approach, J. geol. Soc. , 160( 5), 719– 733. https://doi.org/10.1144/0016-764902-135 Google Scholar CrossRef Search ADS   Barnhart W.D., Lohman R.B., 2013. Phantom earthquakes and triggered aseismic creep: vertical partitioning of strain during earthquake sequences in Iran, Geophys. Res. Lett. , 40( 5), 819– 823. https://doi.org/10.1002/grl.50201 Google Scholar CrossRef Search ADS   Barnhart W.D., Lohman R.B., Mellors R.J., 2013. Active accommodation of plate convergence in Southern Iran: Earthquake locations, triggered aseismic slip, and regional strain rates, J. geophys. Res. , 118( 10), 5699– 5711. https://doi.org/10.1002/jgrb.50380 Google Scholar CrossRef Search ADS   Berberian M., 1995. Master blind thrust faults hidden under the Zagros folds: active basement tectonics and surface morphotectonics, Tectonophysics , 241( 3–4), 193– 224. https://doi.org/10.1016/0040-1951(94)00185-C Google Scholar CrossRef Search ADS   Beutler G., Moore A.W., Mueller I.I., 2009. The international global navigation satellite systems service (IGS): development and achievements, J. Geod. , 83( 3–4), 297– 307. https://doi.org/10.1007/s00190-008-0268-z Google Scholar CrossRef Search ADS   Bird P., 2003. An updated digital model of plate boundaries, Geochem. Geophys. Geosyst. , 4( 3), 1027, doi:10.1029/2001GC000252. https://doi.org/10.1029/2001GC000252 Google Scholar CrossRef Search ADS   Bosold A., Schwarzhans W., Julapour A., Ashrazadeh A.R., Ehsani S.M., 2005. The structural geology of the High Central Zagros revisited (Iran), Petrol. Geosci. , 11( 3), 225– 238. https://doi.org/10.1144/1354-079304-646 Google Scholar CrossRef Search ADS   Clarke P.J. et al.  , 1998. Crustal strain in central Greece from repeated GPS measurements in the interval 1989–1997, Geophys. J. Int. , 135( 1), 195– 214. https://doi.org/10.1046/j.1365-246X.1998.00633.x Google Scholar CrossRef Search ADS   Colman-Sadd S., 1978. Fold development in Zagros simply folded belt, Southwest Iran, AAPG Bull. , 62, 984– 1003. Copley A., Jackson J., 2006. Active tectonics of the Turkish-Iranian Plateau, Tectonics , 25( 6), TC6006, doi:10.1029/2005TC001906. https://doi.org/10.1029/2005TC001906 Google Scholar CrossRef Search ADS   Davies R., England P., Parsons B., Billiris H., Paradissis D., Veis G., 1997. Geodetic strain of Greece in the interval 1892–1992, J. geophys. Res. , 102( B11), 24 571-24 588. https://doi.org/10.1029/97JB01644 Google Scholar CrossRef Search ADS   Djamour Y., Vernant P., Nankali H.R., Tavakoli F., 2011. NW Iran-eastern Turkey present-day kinematics: results from the Iranian permanent GPS network, Earth planet. Sci. Lett. , 307( 1–2), 27– 34. https://doi.org/10.1016/j.epsl.2011.04.029 Google Scholar CrossRef Search ADS   Dziewonski A.M., Chou T.-A., Woodhouse J.H., 1981. Determination of earthquake source parameters from waveform data for studies of global and regional seismicity, J. geophys. Res. , 86( B4), 2825– 2852. https://doi.org/10.1029/JB086iB04p02825 Google Scholar CrossRef Search ADS   Edgell H.S, 1991. Proterozoic salt basins of the Persian Gulf area and their role in hydrocarbon generation, Precambrian Res. , 54( 1), 1– 14. https://doi.org/10.1016/0301-9268(91)90065-I Google Scholar CrossRef Search ADS   Ekström G., Nettles M., Dziewonski A.M., 2012. The global CMT project 2004-2010: centroid-moment tensors for 13,017 earthquakes, Phys. Earth planet. Inter. , 200–201, 1– 9. https://doi.org/10.1016/j.pepi.2012.04.002 Google Scholar CrossRef Search ADS   Elliott J.R. et al.  , 2015. The 2013 Mw 6.2 Khaki-Shonbe (Iran) Earthquake: insights into seismic and aseismic shortening of the Zagros sedimentary cover, Earth Space Sci. , 2( 11), 435– 471. https://doi.org/10.1002/2015EA000098 Google Scholar CrossRef Search ADS   Engdahl R.E., Jackson J.A., Myers S.C., Bergman E.A., Priestley K. 2006. Relocation and assessment of seismicity in the Iran region, Geophys. J. Int. , 167( 2), 761– 778. https://doi.org/10.1111/j.1365-246X.2006.03127.x Google Scholar CrossRef Search ADS   Frohlich C., 1992. Triangle diagrams: ternary graphs to display similarity and diversity of earthquake focal mechanisms, Phys. Earth planet. Inter. , 75( 1–3), 193– 198. https://doi.org/10.1016/0031-9201(92)90130-N Google Scholar CrossRef Search ADS   Ghasemi A., Talbot C.J., 2006. A new tectonic scenario for the Sanandaj-Sirjan Zone (Iran), J. Asian Earth Sci. , 26( 6), 683– 693. https://doi.org/10.1016/j.jseaes.2005.01.003 Google Scholar CrossRef Search ADS   Goldsmith N.F., Martinell J., Demarcq G., Bohn-Havas M., Dockery D.T.III, 1994. Sr-isotopic calibration of Cenozoic bivalvia and Early Miocene migrations: Eurasian Carnivores to Africa (the Hazeva Formation, Israel) and African gazelles and proboscidia to Ipolytarnoc, Hungary, Newslett. Stratig. , 31( 3), 167– 183. https://doi.org/10.1127/nos/31/1994/167 Google Scholar CrossRef Search ADS   Guest B., Stockli D.F., Grove M., Axen G.J., Lam P.S., Hassanzadeh J., 2006. Thermal histories from the central Alborz Mountains, northern Iran: implications for the spatial and temporal distribution of deformation in northern Iran, Bull. geol. Soc. Am. , 118( 11–12), 1507– 1521. https://doi.org/10.1130/B25819.1 Google Scholar CrossRef Search ADS   Haines A.J., Holt W.E., 1993. A procedure for obtaining the complete horizontal motions within zones of distributed deformation from the inversion of strain rate data, J. geophys. Res. , 98( B7), 12 057-12 082. https://doi.org/10.1029/93JB00892 Hanks T.C., Kanamori H., 1979. A moment magnitude scale, J. geophys. Res. , 84( B5), 12057– 12082. https://doi.org/10.1029/JB084iB05p02348 Hatzfeld D., Molnar P., 2010. Comparisons of the kinematics and deep structures of the Zagros and Himalaya and of the Iranian and Tibetan plateaus and geodynamic implications, Rev. Geophys. , 48( 2), RG2005, doi:10.1029/2009RG000304. https://doi.org/10.1029/2009RG000304 Google Scholar CrossRef Search ADS   Herring T.A., King R.W., McClusky S.C., 2010. Introduction to GAMIT/GLOBK, Release 10.4. Massachusetts Institute of Technology, Cambridge, MA, 1– 48. Hessami K., Nilforoushan F., Talbot C.J., 2006. Active deformation within the Zagros Mountains deduced from GPS measurements, J. geol. Soc. , 163( 1), 143– 148. https://doi.org/10.1144/0016-764905-031 Google Scholar CrossRef Search ADS   Hyndman R.D., Weichert D.H., 1983. Seismicity and rates of relative motion on the plate boundaries of Western North America, Geophys. J. Int. , 72( 1), 59– 82. https://doi.org/10.1111/j.1365-246X.1983.tb02804.x Google Scholar CrossRef Search ADS   Jackson J., 1992. Partitioning of strike-slip and convergent motion between Eurasia and Arabia in eastern Turkey and the Caucasus, J. geophys. Res. , 97( B9), 12 471– 12 479. https://doi.org/10.1029/92JB00944 Google Scholar CrossRef Search ADS   Jackson J.A., McKenzie D.P., 1988. The relationship between plate motions and seismic moment tensors, and the rates of active deformation in the Mediterranean and Middle East, Geophys. J. Int. , 93( 1), 45– 73. https://doi.org/10.1111/j.1365-246X.1988.tb01387.x Google Scholar CrossRef Search ADS   Jahani S., Callot J.P., Letouzey J., Frizon De Lamotte D., 2009. The eastern termination of the Zagros fold-and-thrust belt, Iran: structures, evolution, and relationships between salt plugs, folding, and faulting, Tectonics , 28( 6), doi:10.1029/2008TC002418. https://doi.org/10.1029/2008TC002418 Karimiparidari S., Zaré M., Memarian H., Kijko A., 2013. Iranian earthquakes, a uniform catalog with moment magnitudes, J. Seismol. , 17( 3), 897– 911. https://doi.org/10.1007/s10950-013-9360-9 Google Scholar CrossRef Search ADS   Kennett B.L.N., Engdahl E.R., Buland R., 1995. Constraints on seismic velocities in the Earth from traveltimes. Geophys. J. Int. , 122( 1), 108– 124. https://doi.org/10.1111/j.1365-246X.1995.tb03540.x Google Scholar CrossRef Search ADS   Kent P.E., 1979. The emergent Hormuz salt plugs of southern Iran, J Petrol. Geol. , 2( 2), 117– 144. https://doi.org/10.1111/j.1747-5457.1979.tb00698.x Google Scholar CrossRef Search ADS   Khodaverdian A., Zafarani H., Rahimian M., 2015. Long term fault slip rates, distributed deformation rates and forecast of seismicity in the Iranian Plateau, Tectonics , 34( 10), 2190– 2220. https://doi.org/10.1002/2014TC003796 Google Scholar CrossRef Search ADS   Kijko A., Graham G., 1998. Parametric-historic procedure for probabilistic seismic hazard analysis Part I: estimation ofmaximum regionalmagnitude mmax, Pure appl. Geophys. , 154( 3), 413– 442. Google Scholar CrossRef Search ADS   Kijko A., Singh M., 2011. Statistical tools for maximum possible earthquake magnitude estimation, Acta Geophys. , 59( 4), 674– 700. https://doi.org/10.2478/s11600-011-0012-6 Google Scholar CrossRef Search ADS   Kostrov B., 1974. Seismic moment and energy of earthquakes and seismic flow of rock, Izv. Acad. Sci. USSR Phys. Solid Earth , 1, 23– 40. Masson F., Anvari M., Djamour Y., Walpersdorf A., Tavakoli F., Daignieres M., Nankali H., Van Gorp S., 2007. Large-scale velocity field and strain tensor in Iran inferred from GPS measurements: new insight for the present-day deformation pattern within NE Iran, Geophys. J. Int. , 170( 1), 436– 440. https://doi.org/10.1111/j.1365-246X.2007.03477.x Google Scholar CrossRef Search ADS   Masson F., Chéry J., Hatzfeld D., Martinod J., Vernant P., Tavakoli F., Ghafory-Ashtiani M., 2005. Seismic versus aseismic deformation in Iran inferred from earthquakes and geodetic data, Geophys. J. Int. , 160( 1), 217– 226. https://doi.org/10.1111/j.1365-246X.2004.02465.x Google Scholar CrossRef Search ADS   Mazhari S.A., Bea F., Amini S., Ghalamghash J., Molina J.F., Montero P., Scarrow J.H., Williams I.S., 2009. The Eocene bimodal Piranshahr massif of the Sanandaj-Sirjan Zone, NW Iran: a marker of the end of the collision in the Zagros orogen, J. geol. Soc. , 166( 1), 53– 69. https://doi.org/10.1144/0016-76492008-022 Google Scholar CrossRef Search ADS   Mazzotti S., Adams J., 2005. Rates and uncertainties on seismic moment and deformation in eastern Canada, J. geophys. Res. , 110( B9), B09301, doi:10.1029/2004JB003510. https://doi.org/10.1029/2004JB003510 Google Scholar CrossRef Search ADS   Mazzotti S., Leonard L.J., Cassidy J.F., Rogers G.C., Halchuk S., 2011. Seismic hazard in western Canada from GPS strain rates versus earthquake catalog, J. geophys. Res. , 116( B12), B12310, doi:10.1029/2011JB008213. https://doi.org/10.1029/2011JB008213 Google Scholar CrossRef Search ADS   McQuarrie N. 2004. Crustal scale geometry of the Zagros fold-thrust belt, Iran, J. Struct. Geol. , 26( 3), 519– 535. https://doi.org/10.1016/j.jsg.2003.08.009 Google Scholar CrossRef Search ADS   Melville C., 1984. The use of Historical records for seismic assessment, Boll. Geofis. Teor. Appl. , 26, 109– 119. Mirzaei N., Gao M., Chen Y.T., Wang J., 1997. A uniform catalog of earthquakes for seismic hazard assessment in Iran, Acta Seismol. Sin. , 10( 6), 713– 726. https://doi.org/10.1007/s11589-997-0003-5 Google Scholar CrossRef Search ADS   Molinaro M., Zeyen H., Laurencin X., 2005. Lithospheric structure beneath the south-eastern Zagros Mountains, Iran: recent slab break-off?, Terra Nova , 17( 1), 1– 6. https://doi.org/10.1111/j.1365-3121.2004.00575.x Google Scholar CrossRef Search ADS   Mousavi, Z., Walpersdorf A., Walker R.T., Tavakoli F., Pathier E., Nankali H., Nilfouroushan F., Djamour Y., 2013. Global Positioning System constraints on the active tectonics of NE Iran and the South Caspian region, Earth planet. Sci. Lett. , 377–378, 287– 298. https://doi.org/10.1016/j.epsl.2013.07.007 Google Scholar CrossRef Search ADS   Mouthereau F., Lacombe O., Vergés J., 2012. Building the Zagros collisional orogen: timing, strain distribution and the dynamics of Arabia/Eurasia plate convergence, Tectonophysics , 532, 27– 60. https://doi.org/10.1016/j.tecto.2012.01.022 Google Scholar CrossRef Search ADS   Mouthereau F., Tensi J., Bellahsen N., Lacombe O., De Boisgrollier T., Kargar S., 2007. Tertiary sequence of deformation in a thin-skinned/thick-skinned collision belt: the Zagros folded belt (Fars, Iran), Tectonics , 26( 5), Tc5006, doi:10.1029/2007tc002098. https://doi.org/10.1029/2007TC002098 Google Scholar CrossRef Search ADS   Nissen E., Jackson J., Jahani S., Tatar M., 2014. Zagros “phantom earthquakes” reassessed-The interplay of seismicity and deep salt flow in the Simply Folded Belt?, J. geophys. Res. , 119( 4), doi:10.1002/2013JB010796. https://doi.org/10.1002/2013JB010796 Nissen E., Tatar M., Jackson J.A., Allen M.B., 2011. New views on earthquake faulting in the Zagros fold-and-thrust belt of Iran, Geophys. J. Int. , 186( 3), 928– 944. https://doi.org/10.1111/j.1365-246X.2011.05119.x Google Scholar CrossRef Search ADS   Okay A.I., Zattin M., Cavazza W., 2010. Apatite fission-track data for the Miocene Arabia-Eurasia collision, Geology , 38( 1), 35– 38. https://doi.org/10.1130/G30234.1 Google Scholar CrossRef Search ADS   Palano M., 2015. On the present-day crustal stress, strain-rate fields and mantle anisotropy pattern of Italy, Geophys. J. Int. , 200( 2), 969– 985. https://doi.org/10.1093/gji/ggu451 Google Scholar CrossRef Search ADS   Palano M., Imprescia P., Gresta S., 2013. Current stress and strain-rate fields across the Dead Sea Fault System: constraints from seismological data and GPS observations, Earth planet. Sci. Lett. , 369–370, 305– 316. https://doi.org/10.1016/j.epsl.2013.03.043 Google Scholar CrossRef Search ADS   Paul A., Kaviani A., Hatzfeld D., Vergne J., Mokhtari M., 2006. Seismological evidence for crustal-scale thrusting in the Zagros mountain belt (Iran), Geophys. J. Int. , 166( 1), 227– 237. https://doi.org/10.1111/j.1365-246X.2006.02920.x Google Scholar CrossRef Search ADS   Peyret M. et al.  , 2009. Present-day strain distribution across the Minab-Zendan-Palami fault system from dense GPS transects, Geophys. J. Int. , 179( 2), 751– 762. https://doi.org/10.1111/j.1365-246X.2009.04321.x Google Scholar CrossRef Search ADS   Reilinger R. et al.  , 2006. GPS constraints on continental deformation in the Africa-Arabia-Eurasia continental collision zone and implications for the dynamics of plate interactions, J. geophys. Res. , 111( B5), B05411, doi:10.1029/2005JB004051. https://doi.org/10.1029/2005JB004051 Google Scholar CrossRef Search ADS   Sattarzadeh Y., Cosgrove J.W., Vita-Finzi C., 1999. The interplay of faulting and folding during the evolution of the Zagros deformation belt, in Forced Folds and Fractures , Vol. 169, pp. 187– 196, eds, Cosgrove J.W., Ameen M.S., Geol. Soc. Spec. Publ. Google Scholar CrossRef Search ADS   Savage J.C., Simpson R.W., 1997. Surface strain accumulation and the seismic moment tensor, Bull. seism. Soc. Am. , 87, 1345– 1353. Shen Z.-K., Wang M., Zeng Y., Wang F., 2015. Optimal interpolation of spatially discretized geodetic data, Bull. seism. Soc. Am. , 105( 4), 2117– 2127. https://doi.org/10.1785/0120140247 Google Scholar CrossRef Search ADS   Sherkati S., Letouzey J., Frizon de Lamotte D. 2006. Central Zagros fold-thrust belt (Iran): new insights from seismic data, field observation, and sandbox modeling, Tectonics , 25( 4), TC4007, doi:10.1029/2004TC001766. https://doi.org/10.1029/2004TC001766 Google Scholar CrossRef Search ADS   Smith W.H.F., Wessel P., 1990. Gridding with continuous curvature splines in tension, Geophysics , 55( 3), 293– 305. https://doi.org/10.1190/1.1442837 Google Scholar CrossRef Search ADS   Talbot C.J., Alavi M., 1996. The past of a future syntaxis across the Zagros, in Salt Tectonics , Vol. 100, pp. 89– 109, eds, Alsop G.I., Blundell D.J., Davison I., Geol. Soc. Spec. Publ.. Google Scholar CrossRef Search ADS   Talebian M., Jackson J., 2004. A reappraisal of earthquake focal mechanisms and active shortening in the Zagros mountains of Iran, Geophys. J. Int. , 156( 3), 506– 526. https://doi.org/10.1111/j.1365-246X.2004.02092.x Google Scholar CrossRef Search ADS   Tavakoli F. et al.  , 2008. Distribution of the right-lateral strike-slip motion from the Main Recent Fault to the Kazerun Fault System (Zagros, Iran): evidence from present-day GPS velocities, Earth planet. Sci. Lett. , 275( 3–4), 342– 347. https://doi.org/10.1016/j.epsl.2008.08.030 Google Scholar CrossRef Search ADS   Tchernov E., Ginsburg L., Tassy P., Goldsmith N.F., 1987. Miocene mammals of the Negev (Israel), J. Verteb. Paleontol. , 7( 3), 284– 310. https://doi.org/10.1080/02724634.1987.10011661 Google Scholar CrossRef Search ADS   Vernant P., et al., 2004. Contemporary crustal deformation and plate kinematics in middle east constrained by GPS measurement in Iran and northern Oman, Geophys. J. Int. , 157, 381– 398. Google Scholar CrossRef Search ADS   Walpersdorf A., et al., 2006. Difference in the GPS deformation pattern of North and Central Zagros (Iran), Geophys. J. Int. , 167( 3), 1077– 1088. https://doi.org/10.1111/j.1365-246X.2006.03147.x Google Scholar CrossRef Search ADS   Walpersdorf A. et al.  , 2014. Present-day kinematics and fault slip rates in Eastern Iran, derived from 11 years of GPS data, J. geophys. Res. , 119( 2), 1359– 1383. https://doi.org/10.1002/2013JB010620 Google Scholar CrossRef Search ADS   Weichert D.H., 1980. Estimation of the earthquake recurrence parameters from unequal observation periods for different magnitudes, Bull. seism. Soc. Am. , 70, 1337– 1346. Wells D.L., Coppersmith K.J., 1994. New empirical relationship among magnitude, rupture length, rupture width, rupture area, and surface displacement, Bull. seism. Soc. Am. , 84, 974– 1002. Wessel P., Bercovici D., 1998. Interpolation with splines in tension: a Green's function approach, Math. Geol. , 30( 1), 77– 93. https://doi.org/10.1023/A:1021713421882 Google Scholar CrossRef Search ADS   Zarifi Z., Nilfouroushan F., Raeesi M., 2014. Crustal Stress Map of Iran: insight from seismic and geodetic computations, Pure appl. Geophys. , 171( 7), 1219– 1236. https://doi.org/10.1007/s00024-013-0711-9 Google Scholar CrossRef Search ADS   Zoback M.L., 1992. First- and second-order patterns of stress in the lithosphere: the World Stress Map Project, J. geophys. Res. , 97( B8), 11 703– 11 728. https://doi.org/10.1029/92JB00132 Google Scholar CrossRef Search ADS   SUPPORTING INFORMATION Supplementary data are available at GJI online. Figure S1. (a) Strain-rate field according to a Gaussian function for distance weighting and Voronoi cell for areal weighting. (b) Strain-rate field computed with a quadratic function for distance weighting and Voronoi cell for areal weighting. For both computations, we used a weighting threshold value of 24. Strain-rates derived from the velocity interpolation only (the grid cell contain zero or just one GPS site) are reported as white arrows. Uncertainties are also reported. Figure S2. Second invariant from strain-rate fields computed by adopting (a) a ‘spline in tension’ function in deriving a continuous velocity gradient tensor over the study area, (b) a Gaussian function for distance weighting and Voronoi cell for areal weighting, (c) a quadratic function for distance weighting and Voronoi cell for areal weighting. Figure S3. Magnitude histogram plot for earthquakes collected in the database. Figure S4. Time–magnitude plot for earthquakes collected in the database. Figure S5. Depth distribution of earthquakes collected in the database. Figure S6. Examples of cumulative frequency–magnitude distributions (blue diamonds) of earthquakes for selected cells. The red line represents the truncated Gutenberg–Richter function according to eq. (2). Parameters used in eq. (2) are reported in Tables S6 and S8. Figure S7. Observed (in blue; date of the event is also reported) and estimated (in red; associated uncertainties are also reported; see and Table S8 for details) maximum magnitude over the investigated area. Cells coloured in light red have been omitted from the computation (see the main text for details). Figure S8. Moment-rate estimates computed in this study (see Table 1 in the main text for details). GM: geodetic moment-rates computed by adopting a ‘spline in tension’ approach (1) and Gaussian function (2) for distance weighting and Voronoi cell for areal weighting in deriving a continuous velocity gradient tensor over the study area. SM: seismic moment-rates computed by adopting a cumulative truncated Gutenberg–Richter distribution (GR) and a Kostrov summation method (K). Uncertainties (67 per cent confidence interval) are also reported; uncertainties for SM(K) can be considered as ∼5–10 per cent of the estimated value, reflecting a standard error of 0.2 on earthquake magnitudes. Figure S9. Comparison of the seismic coupling coefficient (expressed as a percentage) computed as: SCC(1), ratio between SM(GR) and GM(1) and SCC(2), ratio between SM(GR) and GM(2). Abbreviations are as Fig. S8. Table S1. Historical events used in this study as collected from MIRZ97 (Mirzaei et al.1997), AM&JA98 (Ambraseys & Jackson 1998) and BER95 (Berberian 1995). This table has been provided as text file. Table S2. Site coordinates and velocities (mm yr−1) referred to ITRF2008 for sites used to define the Euler parameters for the fixed Arabian plate. Uncertainties are within the 1σ confidence level. Residual velocities with respect the Arabian plate are also reported. (*) Solutions from ArRajehi et al. (2010). This table has been provided as text file. Table S3. Euler pole parameters, associated errors (3σ) and covariance matrix for the Arabian plate. Equivalent rotation vector Ω (ωx, ωy, ωz) in a Cartesian frame (and associated standard deviation) are as follow: 0.3323 ± 0.0037, −0.0194 ± 0.0036 and 0.4079 ± 0.0028 (deg Myr−1 units). Table S4. Site coordinates, velocities (mm yr−1) referred to ITRF2008 and Arabian plate used in this study. Uncertainties are within the 1σ confidence level. References are also reported. Table S5. Database of instrumental seismicity used in this study. The database has been collected from the ISC online catalogue. Table S6. Summary of earthquake catalogue statistic. NEq, number of earthquakes; Mc, Magnitude of completeness; a and b, seismicity level and slope of the Gutenberg–Richter recurrence relation and associated uncertainty; dT, time interval. This table has been provided as text file. Table S7. Summary of maximum magnitude Mx estimation by applying the 12 procedures implemented in the MMAX toolbox (Kijko & Singh 2011) and by using as input our instrumental database. We reported also the highest magnitude recorded for each cell in the instrumental (see Table S5) and/or historical catalogues (Mirzaei et al.1997; Ambraseys & Jackson 1998; see Table S1 for details). Abbreviations: N-P-G, Non-Parametric with Gaussian kernel; N-P-OS, Non-Parametric procedure based on order statistic; LM, procedure based on few a largest earthquake, nL1, L1 norm regression; LS, Least square approach; T-P, Tate−Pisarenko procedure; K-S_a, Kijko–Sellevol with Cramér's approximation; K-S_e, Kijko–Sellevol (exact solution); T-P-B, Tate–Pisarenko–Bayes procedure; K-S-B, Kijko–Sellevoll–Bayes procedure, R-W, Robson–Whitock, R-W-C, Robson–Whitock–Cooke. Some cells have been omitted because the low number of earthquakes (Z008 and Z017; see Table S6) or because not used for the moment-rate comparisons (Z003, Z004, Z009, Z010, Z018, Z024 and Z030). Table S8. Summary of maximum magnitude Mx estimation by applying the 12 procedures implemented in the MMAX toolbox (Kijko & Singh 2011) and by taking into account a catalogue which include both instrumental (see Table S5) and historical seismic records (see Table S1). We reported also the highest magnitude recorded for each cell in the instrumental (Table S5) and/or historical catalogues (Mirzaei et al.1997; Ambraseys & Jackson 1998; Table S1). Abbreviations are as Table S7. Among the different estimations, for each cell, the Mx value having the smallest uncertainties (cells coloured in grey) have been chosen for the seismic moment-rate estimation. Some cells have been omitted because the low number of earthquakes (Z008 and Z017; see Table S6) or because not used for the moment-rate comparisons (Z003, Z004, Z009, Z010, Z018, Z024 and Z030). Please note: Oxford University Press is not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the paper. © The Author(s) 2017. Published by Oxford University Press on behalf of The Royal Astronomical Society.

### Journal

Geophysical Journal InternationalOxford University Press

Published: Apr 1, 2018

## You’re reading a free preview. Subscribe to read the entire article.

### DeepDyve is your personal research library

It’s your single place to instantly
that matters to you.

over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month ### Explore the DeepDyve Library ### Search Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly ### Organize Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place. ### Access Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals. ### Your journals are on DeepDyve Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more. All the latest content is available, no embargo periods. DeepDyve ### Freelancer DeepDyve ### Pro Price FREE$49/month
\$360/year

Save searches from
PubMed

Create lists to

Export lists, citations